ml/1_segment_dataset_arch_compare_finger_tips.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1 Segment Dataset Architecture Comparisons on Finger Tips"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import collections\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import time\n",
    "import os\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import *\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.models import load_model\n",
    "from tensorflow.keras.utils import plot_model\n",
    "import matplotlib.pyplot as plt\n",
    "pd.set_option('display.max_columns', 500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ['HDF5_USE_FILE_LOCKING'] = 'FALSE'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2.0.0'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Datasets and Clean\n",
    "In this configuration the relevant data set should be loaded from the same folder as the notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('jackson_index_1_joined.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data consists of timestamps from the two hardware devices and a diff between them. When the two hardware data streams were stitched together an effor was made to minimize this diff, but the driver configuration did not easily permit eliminating it. This information is included to understand the accuracy of the data, but will not be used during the training.\n",
    "\n",
    "The time data is followed by the 8 channels from the Myo, this data will be used as input features.\n",
    "\n",
    "This is followed by the 63 positional points from the Leap cameras. These will be used as labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Leap timestamp</th>\n",
       "      <th>timestamp diff</th>\n",
       "      <th>emg timestamp</th>\n",
       "      <th>ch1</th>\n",
       "      <th>ch2</th>\n",
       "      <th>ch3</th>\n",
       "      <th>ch4</th>\n",
       "      <th>ch5</th>\n",
       "      <th>ch6</th>\n",
       "      <th>ch7</th>\n",
       "      <th>ch8</th>\n",
       "      <th>Wrist x</th>\n",
       "      <th>Wrist y</th>\n",
       "      <th>Wrist z</th>\n",
       "      <th>Thumb Proximal x</th>\n",
       "      <th>Thumb Proximal y</th>\n",
       "      <th>Thumb Proximal z</th>\n",
       "      <th>Thumb Intermediate x</th>\n",
       "      <th>Thumb Intermediate y</th>\n",
       "      <th>Thumb Intermediate z</th>\n",
       "      <th>Thumb Distal x</th>\n",
       "      <th>Thumb Distal y</th>\n",
       "      <th>Thumb Distal z</th>\n",
       "      <th>Thumb Tip x</th>\n",
       "      <th>Thumb Tip y</th>\n",
       "      <th>Thumb Tip z</th>\n",
       "      <th>Index Proximal x</th>\n",
       "      <th>Index Proximal y</th>\n",
       "      <th>Index Proximal z</th>\n",
       "      <th>Index Intermediate x</th>\n",
       "      <th>Index Intermediate y</th>\n",
       "      <th>Index Intermediate z</th>\n",
       "      <th>Index Distal x</th>\n",
       "      <th>Index Distal y</th>\n",
       "      <th>Index Distal z</th>\n",
       "      <th>Index Tip x</th>\n",
       "      <th>Index Tip y</th>\n",
       "      <th>Index Tip z</th>\n",
       "      <th>Middle Proximal x</th>\n",
       "      <th>Middle Proximal y</th>\n",
       "      <th>Middle Proximal z</th>\n",
       "      <th>Middle Intermediate x</th>\n",
       "      <th>Middle Intermediate y</th>\n",
       "      <th>Middle Intermediate z</th>\n",
       "      <th>Middle Distal x</th>\n",
       "      <th>Middle Distal y</th>\n",
       "      <th>Middle Distal z</th>\n",
       "      <th>Middle Tip x</th>\n",
       "      <th>Middle Tip y</th>\n",
       "      <th>Middle Tip z</th>\n",
       "      <th>Ring Proximal x</th>\n",
       "      <th>Ring Proximal y</th>\n",
       "      <th>Ring Proximal z</th>\n",
       "      <th>Ring Intermediate x</th>\n",
       "      <th>Ring Intermediate y</th>\n",
       "      <th>Ring Intermediate z</th>\n",
       "      <th>Ring Distal x</th>\n",
       "      <th>Ring Distal y</th>\n",
       "      <th>Ring Distal z</th>\n",
       "      <th>Ring Tip x</th>\n",
       "      <th>Ring Tip y</th>\n",
       "      <th>Ring Tip z</th>\n",
       "      <th>Pinky Proximal x</th>\n",
       "      <th>Pinky Proximal y</th>\n",
       "      <th>Pinky Proximal z</th>\n",
       "      <th>Pinky Intermediate x</th>\n",
       "      <th>Pinky Intermediate y</th>\n",
       "      <th>Pinky Intermediate z</th>\n",
       "      <th>Pinky Distal x</th>\n",
       "      <th>Pinky Distal y</th>\n",
       "      <th>Pinky Distal z</th>\n",
       "      <th>Pinky Tip x</th>\n",
       "      <th>Pinky Tip y</th>\n",
       "      <th>Pinky Tip z</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>0.006472</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>56.0</td>\n",
       "      <td>152.0</td>\n",
       "      <td>318.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>168.0</td>\n",
       "      <td>27.481928</td>\n",
       "      <td>0.635315</td>\n",
       "      <td>44.404579</td>\n",
       "      <td>3.819370</td>\n",
       "      <td>-10.501190</td>\n",
       "      <td>57.583588</td>\n",
       "      <td>-39.021576</td>\n",
       "      <td>-14.662689</td>\n",
       "      <td>37.134148</td>\n",
       "      <td>-69.938087</td>\n",
       "      <td>-15.536011</td>\n",
       "      <td>26.992786</td>\n",
       "      <td>-90.684303</td>\n",
       "      <td>-16.765839</td>\n",
       "      <td>18.792263</td>\n",
       "      <td>-31.480103</td>\n",
       "      <td>3.195557</td>\n",
       "      <td>-11.996174</td>\n",
       "      <td>-48.617424</td>\n",
       "      <td>12.979187</td>\n",
       "      <td>-47.950493</td>\n",
       "      <td>-56.789734</td>\n",
       "      <td>2.440277</td>\n",
       "      <td>-66.780174</td>\n",
       "      <td>-59.230820</td>\n",
       "      <td>-12.140320</td>\n",
       "      <td>-73.670944</td>\n",
       "      <td>-12.269012</td>\n",
       "      <td>5.741394</td>\n",
       "      <td>-20.408253</td>\n",
       "      <td>-26.628906</td>\n",
       "      <td>19.429688</td>\n",
       "      <td>-61.925949</td>\n",
       "      <td>-36.830482</td>\n",
       "      <td>8.192657</td>\n",
       "      <td>-84.433563</td>\n",
       "      <td>-41.600212</td>\n",
       "      <td>-7.619537</td>\n",
       "      <td>-91.437363</td>\n",
       "      <td>9.305349</td>\n",
       "      <td>5.784424</td>\n",
       "      <td>-23.192884</td>\n",
       "      <td>-1.900085</td>\n",
       "      <td>17.207031</td>\n",
       "      <td>-62.730690</td>\n",
       "      <td>-11.677986</td>\n",
       "      <td>6.098206</td>\n",
       "      <td>-84.647568</td>\n",
       "      <td>-17.221012</td>\n",
       "      <td>-9.363892</td>\n",
       "      <td>-91.600197</td>\n",
       "      <td>29.291636</td>\n",
       "      <td>1.544373</td>\n",
       "      <td>-23.824600</td>\n",
       "      <td>22.835503</td>\n",
       "      <td>5.419800</td>\n",
       "      <td>-56.729340</td>\n",
       "      <td>17.429138</td>\n",
       "      <td>2.168213</td>\n",
       "      <td>-74.302849</td>\n",
       "      <td>11.652407</td>\n",
       "      <td>-4.481842</td>\n",
       "      <td>-88.201462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>-0.017444</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>54.0</td>\n",
       "      <td>164.0</td>\n",
       "      <td>301.0</td>\n",
       "      <td>148.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>169.0</td>\n",
       "      <td>26.647976</td>\n",
       "      <td>-0.541893</td>\n",
       "      <td>44.900626</td>\n",
       "      <td>2.603665</td>\n",
       "      <td>-11.599823</td>\n",
       "      <td>57.431037</td>\n",
       "      <td>-39.666034</td>\n",
       "      <td>-14.301682</td>\n",
       "      <td>35.652349</td>\n",
       "      <td>-70.089661</td>\n",
       "      <td>-14.468178</td>\n",
       "      <td>24.132261</td>\n",
       "      <td>-90.578381</td>\n",
       "      <td>-15.055008</td>\n",
       "      <td>15.274604</td>\n",
       "      <td>-31.203072</td>\n",
       "      <td>3.832390</td>\n",
       "      <td>-12.509948</td>\n",
       "      <td>-47.597781</td>\n",
       "      <td>14.169205</td>\n",
       "      <td>-48.651957</td>\n",
       "      <td>-55.505336</td>\n",
       "      <td>3.798073</td>\n",
       "      <td>-67.685915</td>\n",
       "      <td>-57.910363</td>\n",
       "      <td>-10.747452</td>\n",
       "      <td>-74.662590</td>\n",
       "      <td>-11.808361</td>\n",
       "      <td>6.275238</td>\n",
       "      <td>-20.517360</td>\n",
       "      <td>-25.454881</td>\n",
       "      <td>20.207542</td>\n",
       "      <td>-62.191739</td>\n",
       "      <td>-35.418246</td>\n",
       "      <td>9.399536</td>\n",
       "      <td>-85.011818</td>\n",
       "      <td>-40.310819</td>\n",
       "      <td>-6.143776</td>\n",
       "      <td>-92.508541</td>\n",
       "      <td>9.811148</td>\n",
       "      <td>6.081985</td>\n",
       "      <td>-22.903098</td>\n",
       "      <td>-0.553401</td>\n",
       "      <td>17.576523</td>\n",
       "      <td>-62.645160</td>\n",
       "      <td>-9.981847</td>\n",
       "      <td>6.514847</td>\n",
       "      <td>-84.736057</td>\n",
       "      <td>-15.500547</td>\n",
       "      <td>-8.887718</td>\n",
       "      <td>-91.833954</td>\n",
       "      <td>29.747767</td>\n",
       "      <td>1.589165</td>\n",
       "      <td>-23.245819</td>\n",
       "      <td>24.140824</td>\n",
       "      <td>6.374512</td>\n",
       "      <td>-56.173798</td>\n",
       "      <td>19.047623</td>\n",
       "      <td>3.444153</td>\n",
       "      <td>-73.894844</td>\n",
       "      <td>13.431637</td>\n",
       "      <td>-3.077950</td>\n",
       "      <td>-87.919249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>0.003166</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>50.0</td>\n",
       "      <td>161.0</td>\n",
       "      <td>401.0</td>\n",
       "      <td>184.0</td>\n",
       "      <td>49.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>162.0</td>\n",
       "      <td>25.490219</td>\n",
       "      <td>-3.977844</td>\n",
       "      <td>45.407623</td>\n",
       "      <td>0.888691</td>\n",
       "      <td>-15.207642</td>\n",
       "      <td>56.645126</td>\n",
       "      <td>-40.799980</td>\n",
       "      <td>-12.787781</td>\n",
       "      <td>33.687691</td>\n",
       "      <td>-70.586067</td>\n",
       "      <td>-9.926239</td>\n",
       "      <td>20.879833</td>\n",
       "      <td>-91.010010</td>\n",
       "      <td>-7.984314</td>\n",
       "      <td>12.033489</td>\n",
       "      <td>-30.784788</td>\n",
       "      <td>5.359344</td>\n",
       "      <td>-12.988110</td>\n",
       "      <td>-46.170105</td>\n",
       "      <td>17.185913</td>\n",
       "      <td>-49.120426</td>\n",
       "      <td>-53.684822</td>\n",
       "      <td>7.380035</td>\n",
       "      <td>-68.608513</td>\n",
       "      <td>-56.010406</td>\n",
       "      <td>-6.984161</td>\n",
       "      <td>-75.977188</td>\n",
       "      <td>-11.141672</td>\n",
       "      <td>7.823792</td>\n",
       "      <td>-20.362720</td>\n",
       "      <td>-23.704641</td>\n",
       "      <td>23.463989</td>\n",
       "      <td>-61.773483</td>\n",
       "      <td>-33.332466</td>\n",
       "      <td>13.942474</td>\n",
       "      <td>-85.301781</td>\n",
       "      <td>-38.390213</td>\n",
       "      <td>-0.982178</td>\n",
       "      <td>-93.875229</td>\n",
       "      <td>10.523100</td>\n",
       "      <td>7.254333</td>\n",
       "      <td>-22.239563</td>\n",
       "      <td>1.277081</td>\n",
       "      <td>20.302490</td>\n",
       "      <td>-61.781082</td>\n",
       "      <td>-7.774546</td>\n",
       "      <td>10.311188</td>\n",
       "      <td>-84.531990</td>\n",
       "      <td>-13.392775</td>\n",
       "      <td>-4.599304</td>\n",
       "      <td>-92.547852</td>\n",
       "      <td>30.355955</td>\n",
       "      <td>2.308136</td>\n",
       "      <td>-22.387161</td>\n",
       "      <td>26.408840</td>\n",
       "      <td>10.257629</td>\n",
       "      <td>-54.954762</td>\n",
       "      <td>21.679298</td>\n",
       "      <td>8.039429</td>\n",
       "      <td>-72.880737</td>\n",
       "      <td>16.065914</td>\n",
       "      <td>1.318481</td>\n",
       "      <td>-86.812164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>0.010059</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>40.0</td>\n",
       "      <td>170.0</td>\n",
       "      <td>451.0</td>\n",
       "      <td>206.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>25.153648</td>\n",
       "      <td>-5.140533</td>\n",
       "      <td>45.478523</td>\n",
       "      <td>0.395868</td>\n",
       "      <td>-16.426270</td>\n",
       "      <td>56.309441</td>\n",
       "      <td>-41.075424</td>\n",
       "      <td>-11.990051</td>\n",
       "      <td>33.259583</td>\n",
       "      <td>-70.632256</td>\n",
       "      <td>-7.888000</td>\n",
       "      <td>20.259293</td>\n",
       "      <td>-91.043556</td>\n",
       "      <td>-4.950928</td>\n",
       "      <td>11.661678</td>\n",
       "      <td>-30.656673</td>\n",
       "      <td>5.864960</td>\n",
       "      <td>-13.072382</td>\n",
       "      <td>-45.760315</td>\n",
       "      <td>17.912903</td>\n",
       "      <td>-49.250374</td>\n",
       "      <td>-53.120193</td>\n",
       "      <td>8.142944</td>\n",
       "      <td>-68.815437</td>\n",
       "      <td>-55.381058</td>\n",
       "      <td>-6.207825</td>\n",
       "      <td>-76.230270</td>\n",
       "      <td>-10.946850</td>\n",
       "      <td>8.341492</td>\n",
       "      <td>-20.262655</td>\n",
       "      <td>-23.150558</td>\n",
       "      <td>24.345184</td>\n",
       "      <td>-61.641949</td>\n",
       "      <td>-32.622368</td>\n",
       "      <td>15.204285</td>\n",
       "      <td>-85.383560</td>\n",
       "      <td>-37.708450</td>\n",
       "      <td>0.511047</td>\n",
       "      <td>-94.331711</td>\n",
       "      <td>10.725384</td>\n",
       "      <td>7.651855</td>\n",
       "      <td>-22.008507</td>\n",
       "      <td>1.703369</td>\n",
       "      <td>20.909332</td>\n",
       "      <td>-61.532143</td>\n",
       "      <td>-7.254799</td>\n",
       "      <td>11.117615</td>\n",
       "      <td>-84.406494</td>\n",
       "      <td>-12.890709</td>\n",
       "      <td>-3.660126</td>\n",
       "      <td>-92.652481</td>\n",
       "      <td>30.522303</td>\n",
       "      <td>2.563080</td>\n",
       "      <td>-22.131792</td>\n",
       "      <td>27.211058</td>\n",
       "      <td>11.719635</td>\n",
       "      <td>-54.452309</td>\n",
       "      <td>22.549606</td>\n",
       "      <td>9.592224</td>\n",
       "      <td>-72.407104</td>\n",
       "      <td>16.854794</td>\n",
       "      <td>2.514252</td>\n",
       "      <td>-86.127060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>-0.013824</td>\n",
       "      <td>1.571862e+09</td>\n",
       "      <td>45.0</td>\n",
       "      <td>172.0</td>\n",
       "      <td>480.0</td>\n",
       "      <td>212.0</td>\n",
       "      <td>53.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>24.743472</td>\n",
       "      <td>-6.742635</td>\n",
       "      <td>45.491520</td>\n",
       "      <td>-0.110102</td>\n",
       "      <td>-18.394918</td>\n",
       "      <td>55.696049</td>\n",
       "      <td>-41.305864</td>\n",
       "      <td>-12.705556</td>\n",
       "      <td>32.434175</td>\n",
       "      <td>-70.632487</td>\n",
       "      <td>-7.905243</td>\n",
       "      <td>19.158581</td>\n",
       "      <td>-90.982437</td>\n",
       "      <td>-4.420380</td>\n",
       "      <td>10.624847</td>\n",
       "      <td>-30.537937</td>\n",
       "      <td>6.328888</td>\n",
       "      <td>-13.131980</td>\n",
       "      <td>-45.344407</td>\n",
       "      <td>18.750071</td>\n",
       "      <td>-49.306426</td>\n",
       "      <td>-52.401997</td>\n",
       "      <td>9.165934</td>\n",
       "      <td>-69.073558</td>\n",
       "      <td>-54.474183</td>\n",
       "      <td>-5.081034</td>\n",
       "      <td>-76.738945</td>\n",
       "      <td>-10.764413</td>\n",
       "      <td>9.051208</td>\n",
       "      <td>-20.053380</td>\n",
       "      <td>-22.729432</td>\n",
       "      <td>25.211934</td>\n",
       "      <td>-61.441198</td>\n",
       "      <td>-31.890605</td>\n",
       "      <td>16.455200</td>\n",
       "      <td>-85.447795</td>\n",
       "      <td>-36.853535</td>\n",
       "      <td>2.071289</td>\n",
       "      <td>-94.948532</td>\n",
       "      <td>10.922589</td>\n",
       "      <td>8.417521</td>\n",
       "      <td>-21.627883</td>\n",
       "      <td>2.164737</td>\n",
       "      <td>22.093313</td>\n",
       "      <td>-61.068199</td>\n",
       "      <td>-6.513943</td>\n",
       "      <td>12.694092</td>\n",
       "      <td>-84.212934</td>\n",
       "      <td>-12.022625</td>\n",
       "      <td>-1.847493</td>\n",
       "      <td>-92.948497</td>\n",
       "      <td>30.719985</td>\n",
       "      <td>3.331126</td>\n",
       "      <td>-21.752429</td>\n",
       "      <td>28.355796</td>\n",
       "      <td>14.152852</td>\n",
       "      <td>-53.633718</td>\n",
       "      <td>23.850754</td>\n",
       "      <td>12.003235</td>\n",
       "      <td>-71.625646</td>\n",
       "      <td>18.137719</td>\n",
       "      <td>4.223226</td>\n",
       "      <td>-84.948919</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Leap timestamp  timestamp diff  emg timestamp   ch1    ch2    ch3    ch4  \\\n",
       "0    1.571862e+09        0.006472   1.571862e+09  56.0  152.0  318.0   58.0   \n",
       "1    1.571862e+09       -0.017444   1.571862e+09  54.0  164.0  301.0  148.0   \n",
       "2    1.571862e+09        0.003166   1.571862e+09  50.0  161.0  401.0  184.0   \n",
       "3    1.571862e+09        0.010059   1.571862e+09  40.0  170.0  451.0  206.0   \n",
       "4    1.571862e+09       -0.013824   1.571862e+09  45.0  172.0  480.0  212.0   \n",
       "\n",
       "    ch5   ch6   ch7    ch8    Wrist x   Wrist y    Wrist z  Thumb Proximal x  \\\n",
       "0  30.0  26.0  37.0  168.0  27.481928  0.635315  44.404579          3.819370   \n",
       "1  43.0  29.0  36.0  169.0  26.647976 -0.541893  44.900626          2.603665   \n",
       "2  49.0  32.0  37.0  162.0  25.490219 -3.977844  45.407623          0.888691   \n",
       "3  52.0  32.0  28.0   82.0  25.153648 -5.140533  45.478523          0.395868   \n",
       "4  53.0  36.0  41.0  102.0  24.743472 -6.742635  45.491520         -0.110102   \n",
       "\n",
       "   Thumb Proximal y  Thumb Proximal z  Thumb Intermediate x  \\\n",
       "0        -10.501190         57.583588            -39.021576   \n",
       "1        -11.599823         57.431037            -39.666034   \n",
       "2        -15.207642         56.645126            -40.799980   \n",
       "3        -16.426270         56.309441            -41.075424   \n",
       "4        -18.394918         55.696049            -41.305864   \n",
       "\n",
       "   Thumb Intermediate y  Thumb Intermediate z  Thumb Distal x  Thumb Distal y  \\\n",
       "0            -14.662689             37.134148      -69.938087      -15.536011   \n",
       "1            -14.301682             35.652349      -70.089661      -14.468178   \n",
       "2            -12.787781             33.687691      -70.586067       -9.926239   \n",
       "3            -11.990051             33.259583      -70.632256       -7.888000   \n",
       "4            -12.705556             32.434175      -70.632487       -7.905243   \n",
       "\n",
       "   Thumb Distal z  Thumb Tip x  Thumb Tip y  Thumb Tip z  Index Proximal x  \\\n",
       "0       26.992786   -90.684303   -16.765839    18.792263        -31.480103   \n",
       "1       24.132261   -90.578381   -15.055008    15.274604        -31.203072   \n",
       "2       20.879833   -91.010010    -7.984314    12.033489        -30.784788   \n",
       "3       20.259293   -91.043556    -4.950928    11.661678        -30.656673   \n",
       "4       19.158581   -90.982437    -4.420380    10.624847        -30.537937   \n",
       "\n",
       "   Index Proximal y  Index Proximal z  Index Intermediate x  \\\n",
       "0          3.195557        -11.996174            -48.617424   \n",
       "1          3.832390        -12.509948            -47.597781   \n",
       "2          5.359344        -12.988110            -46.170105   \n",
       "3          5.864960        -13.072382            -45.760315   \n",
       "4          6.328888        -13.131980            -45.344407   \n",
       "\n",
       "   Index Intermediate y  Index Intermediate z  Index Distal x  Index Distal y  \\\n",
       "0             12.979187            -47.950493      -56.789734        2.440277   \n",
       "1             14.169205            -48.651957      -55.505336        3.798073   \n",
       "2             17.185913            -49.120426      -53.684822        7.380035   \n",
       "3             17.912903            -49.250374      -53.120193        8.142944   \n",
       "4             18.750071            -49.306426      -52.401997        9.165934   \n",
       "\n",
       "   Index Distal z  Index Tip x  Index Tip y  Index Tip z  Middle Proximal x  \\\n",
       "0      -66.780174   -59.230820   -12.140320   -73.670944         -12.269012   \n",
       "1      -67.685915   -57.910363   -10.747452   -74.662590         -11.808361   \n",
       "2      -68.608513   -56.010406    -6.984161   -75.977188         -11.141672   \n",
       "3      -68.815437   -55.381058    -6.207825   -76.230270         -10.946850   \n",
       "4      -69.073558   -54.474183    -5.081034   -76.738945         -10.764413   \n",
       "\n",
       "   Middle Proximal y  Middle Proximal z  Middle Intermediate x  \\\n",
       "0           5.741394         -20.408253             -26.628906   \n",
       "1           6.275238         -20.517360             -25.454881   \n",
       "2           7.823792         -20.362720             -23.704641   \n",
       "3           8.341492         -20.262655             -23.150558   \n",
       "4           9.051208         -20.053380             -22.729432   \n",
       "\n",
       "   Middle Intermediate y  Middle Intermediate z  Middle Distal x  \\\n",
       "0              19.429688             -61.925949       -36.830482   \n",
       "1              20.207542             -62.191739       -35.418246   \n",
       "2              23.463989             -61.773483       -33.332466   \n",
       "3              24.345184             -61.641949       -32.622368   \n",
       "4              25.211934             -61.441198       -31.890605   \n",
       "\n",
       "   Middle Distal y  Middle Distal z  Middle Tip x  Middle Tip y  Middle Tip z  \\\n",
       "0         8.192657       -84.433563    -41.600212     -7.619537    -91.437363   \n",
       "1         9.399536       -85.011818    -40.310819     -6.143776    -92.508541   \n",
       "2        13.942474       -85.301781    -38.390213     -0.982178    -93.875229   \n",
       "3        15.204285       -85.383560    -37.708450      0.511047    -94.331711   \n",
       "4        16.455200       -85.447795    -36.853535      2.071289    -94.948532   \n",
       "\n",
       "   Ring Proximal x  Ring Proximal y  Ring Proximal z  Ring Intermediate x  \\\n",
       "0         9.305349         5.784424       -23.192884            -1.900085   \n",
       "1         9.811148         6.081985       -22.903098            -0.553401   \n",
       "2        10.523100         7.254333       -22.239563             1.277081   \n",
       "3        10.725384         7.651855       -22.008507             1.703369   \n",
       "4        10.922589         8.417521       -21.627883             2.164737   \n",
       "\n",
       "   Ring Intermediate y  Ring Intermediate z  Ring Distal x  Ring Distal y  \\\n",
       "0            17.207031           -62.730690     -11.677986       6.098206   \n",
       "1            17.576523           -62.645160      -9.981847       6.514847   \n",
       "2            20.302490           -61.781082      -7.774546      10.311188   \n",
       "3            20.909332           -61.532143      -7.254799      11.117615   \n",
       "4            22.093313           -61.068199      -6.513943      12.694092   \n",
       "\n",
       "   Ring Distal z  Ring Tip x  Ring Tip y  Ring Tip z  Pinky Proximal x  \\\n",
       "0     -84.647568  -17.221012   -9.363892  -91.600197         29.291636   \n",
       "1     -84.736057  -15.500547   -8.887718  -91.833954         29.747767   \n",
       "2     -84.531990  -13.392775   -4.599304  -92.547852         30.355955   \n",
       "3     -84.406494  -12.890709   -3.660126  -92.652481         30.522303   \n",
       "4     -84.212934  -12.022625   -1.847493  -92.948497         30.719985   \n",
       "\n",
       "   Pinky Proximal y  Pinky Proximal z  Pinky Intermediate x  \\\n",
       "0          1.544373        -23.824600             22.835503   \n",
       "1          1.589165        -23.245819             24.140824   \n",
       "2          2.308136        -22.387161             26.408840   \n",
       "3          2.563080        -22.131792             27.211058   \n",
       "4          3.331126        -21.752429             28.355796   \n",
       "\n",
       "   Pinky Intermediate y  Pinky Intermediate z  Pinky Distal x  Pinky Distal y  \\\n",
       "0              5.419800            -56.729340       17.429138        2.168213   \n",
       "1              6.374512            -56.173798       19.047623        3.444153   \n",
       "2             10.257629            -54.954762       21.679298        8.039429   \n",
       "3             11.719635            -54.452309       22.549606        9.592224   \n",
       "4             14.152852            -53.633718       23.850754       12.003235   \n",
       "\n",
       "   Pinky Distal z  Pinky Tip x  Pinky Tip y  Pinky Tip z  \n",
       "0      -74.302849    11.652407    -4.481842   -88.201462  \n",
       "1      -73.894844    13.431637    -3.077950   -87.919249  \n",
       "2      -72.880737    16.065914     1.318481   -86.812164  \n",
       "3      -72.407104    16.854794     2.514252   -86.127060  \n",
       "4      -71.625646    18.137719     4.223226   -84.948919  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(labels=[\"Leap timestamp\", \"timestamp diff\", \"emg timestamp\"], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ch1</th>\n",
       "      <th>ch2</th>\n",
       "      <th>ch3</th>\n",
       "      <th>ch4</th>\n",
       "      <th>ch5</th>\n",
       "      <th>ch6</th>\n",
       "      <th>ch7</th>\n",
       "      <th>ch8</th>\n",
       "      <th>Wrist x</th>\n",
       "      <th>Wrist y</th>\n",
       "      <th>Wrist z</th>\n",
       "      <th>Thumb Proximal x</th>\n",
       "      <th>Thumb Proximal y</th>\n",
       "      <th>Thumb Proximal z</th>\n",
       "      <th>Thumb Intermediate x</th>\n",
       "      <th>Thumb Intermediate y</th>\n",
       "      <th>Thumb Intermediate z</th>\n",
       "      <th>Thumb Distal x</th>\n",
       "      <th>Thumb Distal y</th>\n",
       "      <th>Thumb Distal z</th>\n",
       "      <th>Thumb Tip x</th>\n",
       "      <th>Thumb Tip y</th>\n",
       "      <th>Thumb Tip z</th>\n",
       "      <th>Index Proximal x</th>\n",
       "      <th>Index Proximal y</th>\n",
       "      <th>Index Proximal z</th>\n",
       "      <th>Index Intermediate x</th>\n",
       "      <th>Index Intermediate y</th>\n",
       "      <th>Index Intermediate z</th>\n",
       "      <th>Index Distal x</th>\n",
       "      <th>Index Distal y</th>\n",
       "      <th>Index Distal z</th>\n",
       "      <th>Index Tip x</th>\n",
       "      <th>Index Tip y</th>\n",
       "      <th>Index Tip z</th>\n",
       "      <th>Middle Proximal x</th>\n",
       "      <th>Middle Proximal y</th>\n",
       "      <th>Middle Proximal z</th>\n",
       "      <th>Middle Intermediate x</th>\n",
       "      <th>Middle Intermediate y</th>\n",
       "      <th>Middle Intermediate z</th>\n",
       "      <th>Middle Distal x</th>\n",
       "      <th>Middle Distal y</th>\n",
       "      <th>Middle Distal z</th>\n",
       "      <th>Middle Tip x</th>\n",
       "      <th>Middle Tip y</th>\n",
       "      <th>Middle Tip z</th>\n",
       "      <th>Ring Proximal x</th>\n",
       "      <th>Ring Proximal y</th>\n",
       "      <th>Ring Proximal z</th>\n",
       "      <th>Ring Intermediate x</th>\n",
       "      <th>Ring Intermediate y</th>\n",
       "      <th>Ring Intermediate z</th>\n",
       "      <th>Ring Distal x</th>\n",
       "      <th>Ring Distal y</th>\n",
       "      <th>Ring Distal z</th>\n",
       "      <th>Ring Tip x</th>\n",
       "      <th>Ring Tip y</th>\n",
       "      <th>Ring Tip z</th>\n",
       "      <th>Pinky Proximal x</th>\n",
       "      <th>Pinky Proximal y</th>\n",
       "      <th>Pinky Proximal z</th>\n",
       "      <th>Pinky Intermediate x</th>\n",
       "      <th>Pinky Intermediate y</th>\n",
       "      <th>Pinky Intermediate z</th>\n",
       "      <th>Pinky Distal x</th>\n",
       "      <th>Pinky Distal y</th>\n",
       "      <th>Pinky Distal z</th>\n",
       "      <th>Pinky Tip x</th>\n",
       "      <th>Pinky Tip y</th>\n",
       "      <th>Pinky Tip z</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "      <td>87843.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>50.264586</td>\n",
       "      <td>198.888585</td>\n",
       "      <td>393.293205</td>\n",
       "      <td>78.128844</td>\n",
       "      <td>30.925902</td>\n",
       "      <td>24.724622</td>\n",
       "      <td>32.397618</td>\n",
       "      <td>107.234304</td>\n",
       "      <td>3.930715</td>\n",
       "      <td>-11.523523</td>\n",
       "      <td>52.348950</td>\n",
       "      <td>-22.443899</td>\n",
       "      <td>-26.344469</td>\n",
       "      <td>49.591756</td>\n",
       "      <td>-54.614785</td>\n",
       "      <td>-20.539916</td>\n",
       "      <td>13.846868</td>\n",
       "      <td>-69.963748</td>\n",
       "      <td>-17.942858</td>\n",
       "      <td>-15.338769</td>\n",
       "      <td>-88.023557</td>\n",
       "      <td>-14.525735</td>\n",
       "      <td>-28.783127</td>\n",
       "      <td>-23.481550</td>\n",
       "      <td>6.347341</td>\n",
       "      <td>-25.090937</td>\n",
       "      <td>-19.997309</td>\n",
       "      <td>-3.830122</td>\n",
       "      <td>-55.333104</td>\n",
       "      <td>-17.379279</td>\n",
       "      <td>-12.529972</td>\n",
       "      <td>-65.847920</td>\n",
       "      <td>-15.510170</td>\n",
       "      <td>-18.606942</td>\n",
       "      <td>-70.826426</td>\n",
       "      <td>-2.284540</td>\n",
       "      <td>11.292973</td>\n",
       "      <td>-22.424998</td>\n",
       "      <td>6.078287</td>\n",
       "      <td>14.296624</td>\n",
       "      <td>-68.609535</td>\n",
       "      <td>10.934173</td>\n",
       "      <td>12.063061</td>\n",
       "      <td>-95.747102</td>\n",
       "      <td>14.051737</td>\n",
       "      <td>8.567686</td>\n",
       "      <td>-113.280935</td>\n",
       "      <td>18.734271</td>\n",
       "      <td>12.206061</td>\n",
       "      <td>-14.296344</td>\n",
       "      <td>29.419373</td>\n",
       "      <td>20.247313</td>\n",
       "      <td>-55.469134</td>\n",
       "      <td>35.846359</td>\n",
       "      <td>22.651409</td>\n",
       "      <td>-81.444783</td>\n",
       "      <td>40.047243</td>\n",
       "      <td>22.841092</td>\n",
       "      <td>-99.062572</td>\n",
       "      <td>37.854425</td>\n",
       "      <td>8.260404</td>\n",
       "      <td>-6.099804</td>\n",
       "      <td>57.697460</td>\n",
       "      <td>15.985495</td>\n",
       "      <td>-32.908477</td>\n",
       "      <td>68.449561</td>\n",
       "      <td>19.011425</td>\n",
       "      <td>-48.211941</td>\n",
       "      <td>77.724781</td>\n",
       "      <td>20.747419</td>\n",
       "      <td>-61.987531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>16.065377</td>\n",
       "      <td>61.889771</td>\n",
       "      <td>123.922268</td>\n",
       "      <td>31.512907</td>\n",
       "      <td>8.177788</td>\n",
       "      <td>4.985910</td>\n",
       "      <td>13.017049</td>\n",
       "      <td>69.791079</td>\n",
       "      <td>5.672484</td>\n",
       "      <td>8.144951</td>\n",
       "      <td>2.259841</td>\n",
       "      <td>5.918031</td>\n",
       "      <td>7.949196</td>\n",
       "      <td>4.420398</td>\n",
       "      <td>4.894710</td>\n",
       "      <td>5.437073</td>\n",
       "      <td>7.636448</td>\n",
       "      <td>8.430987</td>\n",
       "      <td>9.291401</td>\n",
       "      <td>9.105312</td>\n",
       "      <td>10.886221</td>\n",
       "      <td>12.418942</td>\n",
       "      <td>11.302313</td>\n",
       "      <td>2.659676</td>\n",
       "      <td>4.404711</td>\n",
       "      <td>2.934735</td>\n",
       "      <td>6.755206</td>\n",
       "      <td>29.335207</td>\n",
       "      <td>10.244568</td>\n",
       "      <td>9.356353</td>\n",
       "      <td>44.427175</td>\n",
       "      <td>21.048910</td>\n",
       "      <td>11.165460</td>\n",
       "      <td>54.031633</td>\n",
       "      <td>31.470233</td>\n",
       "      <td>2.602536</td>\n",
       "      <td>3.520150</td>\n",
       "      <td>1.859236</td>\n",
       "      <td>8.639839</td>\n",
       "      <td>9.355284</td>\n",
       "      <td>2.668585</td>\n",
       "      <td>12.125271</td>\n",
       "      <td>14.331292</td>\n",
       "      <td>3.844615</td>\n",
       "      <td>14.346593</td>\n",
       "      <td>17.904034</td>\n",
       "      <td>5.031069</td>\n",
       "      <td>1.957883</td>\n",
       "      <td>2.718714</td>\n",
       "      <td>2.506316</td>\n",
       "      <td>7.988899</td>\n",
       "      <td>9.584357</td>\n",
       "      <td>3.777613</td>\n",
       "      <td>11.780185</td>\n",
       "      <td>14.753008</td>\n",
       "      <td>4.995936</td>\n",
       "      <td>14.318904</td>\n",
       "      <td>18.111585</td>\n",
       "      <td>6.020476</td>\n",
       "      <td>1.189455</td>\n",
       "      <td>3.285129</td>\n",
       "      <td>4.007710</td>\n",
       "      <td>5.691392</td>\n",
       "      <td>8.792838</td>\n",
       "      <td>6.452737</td>\n",
       "      <td>8.286797</td>\n",
       "      <td>12.380805</td>\n",
       "      <td>7.948119</td>\n",
       "      <td>10.596036</td>\n",
       "      <td>15.517381</td>\n",
       "      <td>9.297821</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-12.546410</td>\n",
       "      <td>-39.911087</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-37.846565</td>\n",
       "      <td>-56.398163</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-64.980680</td>\n",
       "      <td>-42.962891</td>\n",
       "      <td>-10.776787</td>\n",
       "      <td>-85.509727</td>\n",
       "      <td>-55.793365</td>\n",
       "      <td>-44.110306</td>\n",
       "      <td>-106.816956</td>\n",
       "      <td>-65.868450</td>\n",
       "      <td>-65.092405</td>\n",
       "      <td>-33.983571</td>\n",
       "      <td>-17.390320</td>\n",
       "      <td>-32.209695</td>\n",
       "      <td>-57.944725</td>\n",
       "      <td>-51.231323</td>\n",
       "      <td>-72.944334</td>\n",
       "      <td>-66.356552</td>\n",
       "      <td>-74.435822</td>\n",
       "      <td>-96.285278</td>\n",
       "      <td>-75.462486</td>\n",
       "      <td>-91.046356</td>\n",
       "      <td>-112.788712</td>\n",
       "      <td>-18.186447</td>\n",
       "      <td>-9.201309</td>\n",
       "      <td>-25.658291</td>\n",
       "      <td>-43.688686</td>\n",
       "      <td>-47.724792</td>\n",
       "      <td>-73.316345</td>\n",
       "      <td>-41.276627</td>\n",
       "      <td>-75.498001</td>\n",
       "      <td>-101.111732</td>\n",
       "      <td>-47.323700</td>\n",
       "      <td>-93.674927</td>\n",
       "      <td>-119.497849</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-1.255035</td>\n",
       "      <td>-23.192884</td>\n",
       "      <td>-19.993179</td>\n",
       "      <td>-34.883713</td>\n",
       "      <td>-65.039101</td>\n",
       "      <td>-20.938644</td>\n",
       "      <td>-61.838303</td>\n",
       "      <td>-92.492859</td>\n",
       "      <td>-21.698006</td>\n",
       "      <td>-80.133446</td>\n",
       "      <td>-111.139832</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-3.554070</td>\n",
       "      <td>-29.067459</td>\n",
       "      <td>-3.880405</td>\n",
       "      <td>-23.929962</td>\n",
       "      <td>-57.870079</td>\n",
       "      <td>-8.428043</td>\n",
       "      <td>-41.576996</td>\n",
       "      <td>-75.311195</td>\n",
       "      <td>-9.027550</td>\n",
       "      <td>-57.863800</td>\n",
       "      <td>-92.172951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>161.000000</td>\n",
       "      <td>319.000000</td>\n",
       "      <td>58.000000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>42.000000</td>\n",
       "      <td>-0.242438</td>\n",
       "      <td>-17.779761</td>\n",
       "      <td>51.269873</td>\n",
       "      <td>-26.955276</td>\n",
       "      <td>-31.974266</td>\n",
       "      <td>46.666297</td>\n",
       "      <td>-57.565168</td>\n",
       "      <td>-24.054184</td>\n",
       "      <td>8.609134</td>\n",
       "      <td>-74.979048</td>\n",
       "      <td>-23.146787</td>\n",
       "      <td>-21.273879</td>\n",
       "      <td>-94.359531</td>\n",
       "      <td>-21.414757</td>\n",
       "      <td>-35.163224</td>\n",
       "      <td>-25.306755</td>\n",
       "      <td>3.487362</td>\n",
       "      <td>-27.082337</td>\n",
       "      <td>-23.896370</td>\n",
       "      <td>-32.369916</td>\n",
       "      <td>-65.043694</td>\n",
       "      <td>-22.947248</td>\n",
       "      <td>-55.228432</td>\n",
       "      <td>-86.745473</td>\n",
       "      <td>-21.924920</td>\n",
       "      <td>-70.320314</td>\n",
       "      <td>-102.752209</td>\n",
       "      <td>-4.031650</td>\n",
       "      <td>9.026708</td>\n",
       "      <td>-23.843467</td>\n",
       "      <td>0.727572</td>\n",
       "      <td>10.076836</td>\n",
       "      <td>-70.561451</td>\n",
       "      <td>3.484806</td>\n",
       "      <td>5.311445</td>\n",
       "      <td>-98.320976</td>\n",
       "      <td>5.274194</td>\n",
       "      <td>0.166153</td>\n",
       "      <td>-116.324841</td>\n",
       "      <td>17.474185</td>\n",
       "      <td>10.479034</td>\n",
       "      <td>-16.125963</td>\n",
       "      <td>24.011989</td>\n",
       "      <td>15.167759</td>\n",
       "      <td>-57.883539</td>\n",
       "      <td>27.745125</td>\n",
       "      <td>15.390785</td>\n",
       "      <td>-84.707203</td>\n",
       "      <td>30.120822</td>\n",
       "      <td>14.094311</td>\n",
       "      <td>-102.988324</td>\n",
       "      <td>37.247028</td>\n",
       "      <td>5.834770</td>\n",
       "      <td>-8.491966</td>\n",
       "      <td>54.367672</td>\n",
       "      <td>9.928127</td>\n",
       "      <td>-36.788572</td>\n",
       "      <td>63.591489</td>\n",
       "      <td>10.774315</td>\n",
       "      <td>-53.140583</td>\n",
       "      <td>71.505948</td>\n",
       "      <td>10.600433</td>\n",
       "      <td>-67.974457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>197.000000</td>\n",
       "      <td>400.000000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>105.000000</td>\n",
       "      <td>4.434735</td>\n",
       "      <td>-11.372952</td>\n",
       "      <td>53.012089</td>\n",
       "      <td>-22.115429</td>\n",
       "      <td>-26.762573</td>\n",
       "      <td>50.422907</td>\n",
       "      <td>-55.824723</td>\n",
       "      <td>-20.496368</td>\n",
       "      <td>14.288343</td>\n",
       "      <td>-72.364769</td>\n",
       "      <td>-18.341064</td>\n",
       "      <td>-15.050502</td>\n",
       "      <td>-91.070063</td>\n",
       "      <td>-14.909607</td>\n",
       "      <td>-28.470984</td>\n",
       "      <td>-23.840214</td>\n",
       "      <td>6.378564</td>\n",
       "      <td>-25.426472</td>\n",
       "      <td>-18.916653</td>\n",
       "      <td>-9.581299</td>\n",
       "      <td>-54.926171</td>\n",
       "      <td>-15.365816</td>\n",
       "      <td>-27.051392</td>\n",
       "      <td>-69.631195</td>\n",
       "      <td>-13.354689</td>\n",
       "      <td>-38.987976</td>\n",
       "      <td>-79.716522</td>\n",
       "      <td>-2.500637</td>\n",
       "      <td>11.299011</td>\n",
       "      <td>-22.744263</td>\n",
       "      <td>5.270168</td>\n",
       "      <td>14.785599</td>\n",
       "      <td>-69.013494</td>\n",
       "      <td>9.963341</td>\n",
       "      <td>12.845184</td>\n",
       "      <td>-96.553566</td>\n",
       "      <td>13.080652</td>\n",
       "      <td>9.603500</td>\n",
       "      <td>-114.442656</td>\n",
       "      <td>18.784060</td>\n",
       "      <td>12.412906</td>\n",
       "      <td>-14.613932</td>\n",
       "      <td>28.324614</td>\n",
       "      <td>20.511307</td>\n",
       "      <td>-55.958656</td>\n",
       "      <td>34.224987</td>\n",
       "      <td>23.243790</td>\n",
       "      <td>-82.324387</td>\n",
       "      <td>37.986172</td>\n",
       "      <td>23.741577</td>\n",
       "      <td>-100.324814</td>\n",
       "      <td>38.018920</td>\n",
       "      <td>8.366096</td>\n",
       "      <td>-6.357288</td>\n",
       "      <td>57.488644</td>\n",
       "      <td>16.321121</td>\n",
       "      <td>-33.525398</td>\n",
       "      <td>68.005737</td>\n",
       "      <td>19.541687</td>\n",
       "      <td>-48.928640</td>\n",
       "      <td>77.124802</td>\n",
       "      <td>21.381683</td>\n",
       "      <td>-62.847455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>59.000000</td>\n",
       "      <td>236.000000</td>\n",
       "      <td>475.000000</td>\n",
       "      <td>89.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>26.000000</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>149.000000</td>\n",
       "      <td>7.507035</td>\n",
       "      <td>-6.124988</td>\n",
       "      <td>54.030175</td>\n",
       "      <td>-18.614227</td>\n",
       "      <td>-20.994272</td>\n",
       "      <td>52.939123</td>\n",
       "      <td>-53.007000</td>\n",
       "      <td>-16.930725</td>\n",
       "      <td>18.984735</td>\n",
       "      <td>-67.658801</td>\n",
       "      <td>-12.110699</td>\n",
       "      <td>-9.918048</td>\n",
       "      <td>-85.059692</td>\n",
       "      <td>-6.516609</td>\n",
       "      <td>-22.585935</td>\n",
       "      <td>-21.538252</td>\n",
       "      <td>9.636242</td>\n",
       "      <td>-23.724293</td>\n",
       "      <td>-15.535723</td>\n",
       "      <td>19.167430</td>\n",
       "      <td>-46.365857</td>\n",
       "      <td>-11.224405</td>\n",
       "      <td>22.916527</td>\n",
       "      <td>-45.162505</td>\n",
       "      <td>-8.025871</td>\n",
       "      <td>24.752247</td>\n",
       "      <td>-39.756063</td>\n",
       "      <td>-0.306380</td>\n",
       "      <td>14.015352</td>\n",
       "      <td>-21.227091</td>\n",
       "      <td>12.753001</td>\n",
       "      <td>19.713528</td>\n",
       "      <td>-67.081800</td>\n",
       "      <td>20.295967</td>\n",
       "      <td>20.464413</td>\n",
       "      <td>-94.214500</td>\n",
       "      <td>25.054950</td>\n",
       "      <td>19.063240</td>\n",
       "      <td>-111.871563</td>\n",
       "      <td>20.237023</td>\n",
       "      <td>14.020319</td>\n",
       "      <td>-12.680142</td>\n",
       "      <td>36.228395</td>\n",
       "      <td>26.828938</td>\n",
       "      <td>-53.549673</td>\n",
       "      <td>45.872225</td>\n",
       "      <td>32.405930</td>\n",
       "      <td>-78.986931</td>\n",
       "      <td>52.240708</td>\n",
       "      <td>34.667305</td>\n",
       "      <td>-96.254533</td>\n",
       "      <td>38.723204</td>\n",
       "      <td>10.615707</td>\n",
       "      <td>-3.641209</td>\n",
       "      <td>62.504412</td>\n",
       "      <td>22.335171</td>\n",
       "      <td>-28.667228</td>\n",
       "      <td>75.451366</td>\n",
       "      <td>27.754917</td>\n",
       "      <td>-42.981016</td>\n",
       "      <td>86.638632</td>\n",
       "      <td>31.554276</td>\n",
       "      <td>-55.772030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>462.000000</td>\n",
       "      <td>1332.000000</td>\n",
       "      <td>1055.000000</td>\n",
       "      <td>1197.000000</td>\n",
       "      <td>806.000000</td>\n",
       "      <td>247.000000</td>\n",
       "      <td>381.000000</td>\n",
       "      <td>1634.000000</td>\n",
       "      <td>44.037949</td>\n",
       "      <td>31.429001</td>\n",
       "      <td>54.706158</td>\n",
       "      <td>24.735947</td>\n",
       "      <td>17.392929</td>\n",
       "      <td>60.297832</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>9.029129</td>\n",
       "      <td>41.423584</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>20.109650</td>\n",
       "      <td>26.992786</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>31.477844</td>\n",
       "      <td>26.328176</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>20.484436</td>\n",
       "      <td>13.875271</td>\n",
       "      <td>4.325998</td>\n",
       "      <td>55.119568</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>16.311575</td>\n",
       "      <td>75.096161</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>24.526445</td>\n",
       "      <td>88.150208</td>\n",
       "      <td>1.552776</td>\n",
       "      <td>4.994682</td>\n",
       "      <td>22.330582</td>\n",
       "      <td>4.815260</td>\n",
       "      <td>27.125025</td>\n",
       "      <td>46.559006</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>39.786163</td>\n",
       "      <td>64.413269</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>47.941965</td>\n",
       "      <td>76.050323</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>23.418606</td>\n",
       "      <td>26.008987</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>49.500938</td>\n",
       "      <td>47.193054</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>65.121246</td>\n",
       "      <td>65.080261</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>75.358543</td>\n",
       "      <td>77.020050</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>39.577545</td>\n",
       "      <td>29.680553</td>\n",
       "      <td>5.625256</td>\n",
       "      <td>68.257573</td>\n",
       "      <td>41.837326</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>83.923373</td>\n",
       "      <td>54.915054</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>97.745659</td>\n",
       "      <td>66.860889</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                ch1           ch2           ch3           ch4           ch5  \\\n",
       "count  87843.000000  87843.000000  87843.000000  87843.000000  87843.000000   \n",
       "mean      50.264586    198.888585    393.293205     78.128844     30.925902   \n",
       "std       16.065377     61.889771    123.922268     31.512907      8.177788   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%       40.000000    161.000000    319.000000     58.000000     27.000000   \n",
       "50%       48.000000    197.000000    400.000000     71.000000     30.000000   \n",
       "75%       59.000000    236.000000    475.000000     89.000000     34.000000   \n",
       "max      462.000000   1332.000000   1055.000000   1197.000000    806.000000   \n",
       "\n",
       "                ch6           ch7           ch8       Wrist x       Wrist y  \\\n",
       "count  87843.000000  87843.000000  87843.000000  87843.000000  87843.000000   \n",
       "mean      24.724622     32.397618    107.234304      3.930715    -11.523523   \n",
       "std        4.985910     13.017049     69.791079      5.672484      8.144951   \n",
       "min        0.000000      0.000000      0.000000    -12.546410    -39.911087   \n",
       "25%       22.000000     25.000000     42.000000     -0.242438    -17.779761   \n",
       "50%       24.000000     30.000000    105.000000      4.434735    -11.372952   \n",
       "75%       26.000000     36.000000    149.000000      7.507035     -6.124988   \n",
       "max      247.000000    381.000000   1634.000000     44.037949     31.429001   \n",
       "\n",
       "            Wrist z  Thumb Proximal x  Thumb Proximal y  Thumb Proximal z  \\\n",
       "count  87843.000000      87843.000000      87843.000000      87843.000000   \n",
       "mean      52.348950        -22.443899        -26.344469         49.591756   \n",
       "std        2.259841          5.918031          7.949196          4.420398   \n",
       "min        0.000000        -37.846565        -56.398163          0.000000   \n",
       "25%       51.269873        -26.955276        -31.974266         46.666297   \n",
       "50%       53.012089        -22.115429        -26.762573         50.422907   \n",
       "75%       54.030175        -18.614227        -20.994272         52.939123   \n",
       "max       54.706158         24.735947         17.392929         60.297832   \n",
       "\n",
       "       Thumb Intermediate x  Thumb Intermediate y  Thumb Intermediate z  \\\n",
       "count          87843.000000          87843.000000          87843.000000   \n",
       "mean             -54.614785            -20.539916             13.846868   \n",
       "std                4.894710              5.437073              7.636448   \n",
       "min              -64.980680            -42.962891            -10.776787   \n",
       "25%              -57.565168            -24.054184              8.609134   \n",
       "50%              -55.824723            -20.496368             14.288343   \n",
       "75%              -53.007000            -16.930725             18.984735   \n",
       "max                0.000000              9.029129             41.423584   \n",
       "\n",
       "       Thumb Distal x  Thumb Distal y  Thumb Distal z   Thumb Tip x  \\\n",
       "count    87843.000000    87843.000000    87843.000000  87843.000000   \n",
       "mean       -69.963748      -17.942858      -15.338769    -88.023557   \n",
       "std          8.430987        9.291401        9.105312     10.886221   \n",
       "min        -85.509727      -55.793365      -44.110306   -106.816956   \n",
       "25%        -74.979048      -23.146787      -21.273879    -94.359531   \n",
       "50%        -72.364769      -18.341064      -15.050502    -91.070063   \n",
       "75%        -67.658801      -12.110699       -9.918048    -85.059692   \n",
       "max          0.000000       20.109650       26.992786      0.000000   \n",
       "\n",
       "        Thumb Tip y   Thumb Tip z  Index Proximal x  Index Proximal y  \\\n",
       "count  87843.000000  87843.000000      87843.000000      87843.000000   \n",
       "mean     -14.525735    -28.783127        -23.481550          6.347341   \n",
       "std       12.418942     11.302313          2.659676          4.404711   \n",
       "min      -65.868450    -65.092405        -33.983571        -17.390320   \n",
       "25%      -21.414757    -35.163224        -25.306755          3.487362   \n",
       "50%      -14.909607    -28.470984        -23.840214          6.378564   \n",
       "75%       -6.516609    -22.585935        -21.538252          9.636242   \n",
       "max       31.477844     26.328176          0.000000         20.484436   \n",
       "\n",
       "       Index Proximal z  Index Intermediate x  Index Intermediate y  \\\n",
       "count      87843.000000          87843.000000          87843.000000   \n",
       "mean         -25.090937            -19.997309             -3.830122   \n",
       "std            2.934735              6.755206             29.335207   \n",
       "min          -32.209695            -57.944725            -51.231323   \n",
       "25%          -27.082337            -23.896370            -32.369916   \n",
       "50%          -25.426472            -18.916653             -9.581299   \n",
       "75%          -23.724293            -15.535723             19.167430   \n",
       "max           13.875271              4.325998             55.119568   \n",
       "\n",
       "       Index Intermediate z  Index Distal x  Index Distal y  Index Distal z  \\\n",
       "count          87843.000000    87843.000000    87843.000000    87843.000000   \n",
       "mean             -55.333104      -17.379279      -12.529972      -65.847920   \n",
       "std               10.244568        9.356353       44.427175       21.048910   \n",
       "min              -72.944334      -66.356552      -74.435822      -96.285278   \n",
       "25%              -65.043694      -22.947248      -55.228432      -86.745473   \n",
       "50%              -54.926171      -15.365816      -27.051392      -69.631195   \n",
       "75%              -46.365857      -11.224405       22.916527      -45.162505   \n",
       "max                0.000000       16.311575       75.096161        0.000000   \n",
       "\n",
       "        Index Tip x   Index Tip y   Index Tip z  Middle Proximal x  \\\n",
       "count  87843.000000  87843.000000  87843.000000       87843.000000   \n",
       "mean     -15.510170    -18.606942    -70.826426          -2.284540   \n",
       "std       11.165460     54.031633     31.470233           2.602536   \n",
       "min      -75.462486    -91.046356   -112.788712         -18.186447   \n",
       "25%      -21.924920    -70.320314   -102.752209          -4.031650   \n",
       "50%      -13.354689    -38.987976    -79.716522          -2.500637   \n",
       "75%       -8.025871     24.752247    -39.756063          -0.306380   \n",
       "max       24.526445     88.150208      1.552776           4.994682   \n",
       "\n",
       "       Middle Proximal y  Middle Proximal z  Middle Intermediate x  \\\n",
       "count       87843.000000       87843.000000           87843.000000   \n",
       "mean           11.292973         -22.424998               6.078287   \n",
       "std             3.520150           1.859236               8.639839   \n",
       "min            -9.201309         -25.658291             -43.688686   \n",
       "25%             9.026708         -23.843467               0.727572   \n",
       "50%            11.299011         -22.744263               5.270168   \n",
       "75%            14.015352         -21.227091              12.753001   \n",
       "max            22.330582           4.815260              27.125025   \n",
       "\n",
       "       Middle Intermediate y  Middle Intermediate z  Middle Distal x  \\\n",
       "count           87843.000000           87843.000000     87843.000000   \n",
       "mean               14.296624             -68.609535        10.934173   \n",
       "std                 9.355284               2.668585        12.125271   \n",
       "min               -47.724792             -73.316345       -41.276627   \n",
       "25%                10.076836             -70.561451         3.484806   \n",
       "50%                14.785599             -69.013494         9.963341   \n",
       "75%                19.713528             -67.081800        20.295967   \n",
       "max                46.559006               0.000000        39.786163   \n",
       "\n",
       "       Middle Distal y  Middle Distal z  Middle Tip x  Middle Tip y  \\\n",
       "count     87843.000000     87843.000000  87843.000000  87843.000000   \n",
       "mean         12.063061       -95.747102     14.051737      8.567686   \n",
       "std          14.331292         3.844615     14.346593     17.904034   \n",
       "min         -75.498001      -101.111732    -47.323700    -93.674927   \n",
       "25%           5.311445       -98.320976      5.274194      0.166153   \n",
       "50%          12.845184       -96.553566     13.080652      9.603500   \n",
       "75%          20.464413       -94.214500     25.054950     19.063240   \n",
       "max          64.413269         0.000000     47.941965     76.050323   \n",
       "\n",
       "       Middle Tip z  Ring Proximal x  Ring Proximal y  Ring Proximal z  \\\n",
       "count  87843.000000     87843.000000     87843.000000     87843.000000   \n",
       "mean    -113.280935        18.734271        12.206061       -14.296344   \n",
       "std        5.031069         1.957883         2.718714         2.506316   \n",
       "min     -119.497849         0.000000        -1.255035       -23.192884   \n",
       "25%     -116.324841        17.474185        10.479034       -16.125963   \n",
       "50%     -114.442656        18.784060        12.412906       -14.613932   \n",
       "75%     -111.871563        20.237023        14.020319       -12.680142   \n",
       "max        0.000000        23.418606        26.008987         0.000000   \n",
       "\n",
       "       Ring Intermediate x  Ring Intermediate y  Ring Intermediate z  \\\n",
       "count         87843.000000         87843.000000         87843.000000   \n",
       "mean             29.419373            20.247313           -55.469134   \n",
       "std               7.988899             9.584357             3.777613   \n",
       "min             -19.993179           -34.883713           -65.039101   \n",
       "25%              24.011989            15.167759           -57.883539   \n",
       "50%              28.324614            20.511307           -55.958656   \n",
       "75%              36.228395            26.828938           -53.549673   \n",
       "max              49.500938            47.193054             0.000000   \n",
       "\n",
       "       Ring Distal x  Ring Distal y  Ring Distal z    Ring Tip x  \\\n",
       "count   87843.000000   87843.000000   87843.000000  87843.000000   \n",
       "mean       35.846359      22.651409     -81.444783     40.047243   \n",
       "std        11.780185      14.753008       4.995936     14.318904   \n",
       "min       -20.938644     -61.838303     -92.492859    -21.698006   \n",
       "25%        27.745125      15.390785     -84.707203     30.120822   \n",
       "50%        34.224987      23.243790     -82.324387     37.986172   \n",
       "75%        45.872225      32.405930     -78.986931     52.240708   \n",
       "max        65.121246      65.080261       0.000000     75.358543   \n",
       "\n",
       "         Ring Tip y    Ring Tip z  Pinky Proximal x  Pinky Proximal y  \\\n",
       "count  87843.000000  87843.000000      87843.000000      87843.000000   \n",
       "mean      22.841092    -99.062572         37.854425          8.260404   \n",
       "std       18.111585      6.020476          1.189455          3.285129   \n",
       "min      -80.133446   -111.139832          0.000000         -3.554070   \n",
       "25%       14.094311   -102.988324         37.247028          5.834770   \n",
       "50%       23.741577   -100.324814         38.018920          8.366096   \n",
       "75%       34.667305    -96.254533         38.723204         10.615707   \n",
       "max       77.020050      0.000000         39.577545         29.680553   \n",
       "\n",
       "       Pinky Proximal z  Pinky Intermediate x  Pinky Intermediate y  \\\n",
       "count      87843.000000          87843.000000          87843.000000   \n",
       "mean          -6.099804             57.697460             15.985495   \n",
       "std            4.007710              5.691392              8.792838   \n",
       "min          -29.067459             -3.880405            -23.929962   \n",
       "25%           -8.491966             54.367672              9.928127   \n",
       "50%           -6.357288             57.488644             16.321121   \n",
       "75%           -3.641209             62.504412             22.335171   \n",
       "max            5.625256             68.257573             41.837326   \n",
       "\n",
       "       Pinky Intermediate z  Pinky Distal x  Pinky Distal y  Pinky Distal z  \\\n",
       "count          87843.000000    87843.000000    87843.000000    87843.000000   \n",
       "mean             -32.908477       68.449561       19.011425      -48.211941   \n",
       "std                6.452737        8.286797       12.380805        7.948119   \n",
       "min              -57.870079       -8.428043      -41.576996      -75.311195   \n",
       "25%              -36.788572       63.591489       10.774315      -53.140583   \n",
       "50%              -33.525398       68.005737       19.541687      -48.928640   \n",
       "75%              -28.667228       75.451366       27.754917      -42.981016   \n",
       "max                0.000000       83.923373       54.915054        0.000000   \n",
       "\n",
       "        Pinky Tip x   Pinky Tip y   Pinky Tip z  \n",
       "count  87843.000000  87843.000000  87843.000000  \n",
       "mean      77.724781     20.747419    -61.987531  \n",
       "std       10.596036     15.517381      9.297821  \n",
       "min       -9.027550    -57.863800    -92.172951  \n",
       "25%       71.505948     10.600433    -67.974457  \n",
       "50%       77.124802     21.381683    -62.847455  \n",
       "75%       86.638632     31.554276    -55.772030  \n",
       "max       97.745659     66.860889      0.000000  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_ar = df.loc[:, 'ch1':'ch8'].values\n",
    "label_ar = df.loc[:, 'Wrist x':].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(87843, 8)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_ar.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(87843, 63)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_ar.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the above evaluation the initial untrained loss is around 2500.\n",
    "Now train the model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(87812, 32, 8)\n",
      "(87812, 63)\n"
     ]
    }
   ],
   "source": [
    "seq_length = 32\n",
    "\n",
    "def overlap_samples(seq_length, feats, labels):\n",
    "    new_l = labels[seq_length - 1:]\n",
    "    feat_list = [feats[i:i + seq_length] for i in range(feats.shape[0] - seq_length + 1)]\n",
    "    new_f = np.array(feat_list)\n",
    "    return new_f, new_l\n",
    "\n",
    "features, labels = overlap_samples(seq_length, feature_ar, label_ar)\n",
    "print(features.shape)\n",
    "print(labels.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training the Model On Tips Only\n",
    "In an attempt to improve the accuracy of prediction of the fingertips the above models are modified and trained only using the fingertip position data to understand if this simplification yields an improvement. This is done on both of the architectures tested above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(87843, 18)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_ar_t = label_ar[:,[0,1,2,12,13,14,24,25,26,36,37,38,48,49,50,60,61,62]]\n",
    "label_ar_t.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(87812, 18)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels_tip = labels[:,[0,1,2,12,13,14,24,25,26,36,37,38,48,49,50,60,61,62]]\n",
    "labels_tip.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tips only with FC Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 70249 samples, validate on 17563 samples\n",
      "Epoch 1/8\n",
      "70249/70249 [==============================] - 33s 472us/sample - loss: 3093.2713 - val_loss: 2894.2592\n",
      "Epoch 2/8\n",
      "70249/70249 [==============================] - 25s 355us/sample - loss: 2336.5584 - val_loss: 1746.8642\n",
      "Epoch 3/8\n",
      "70249/70249 [==============================] - 25s 359us/sample - loss: 1251.9868 - val_loss: 785.9168\n",
      "Epoch 4/8\n",
      "70249/70249 [==============================] - 26s 363us/sample - loss: 909.9474 - val_loss: 486.4390\n",
      "Epoch 5/8\n",
      "70249/70249 [==============================] - 25s 362us/sample - loss: 589.8124 - val_loss: 402.1733\n",
      "Epoch 6/8\n",
      "70249/70249 [==============================] - 26s 367us/sample - loss: 660.9518 - val_loss: 549.6040\n",
      "Epoch 7/8\n",
      "70249/70249 [==============================] - 25s 356us/sample - loss: 529.4689 - val_loss: 454.9528\n",
      "Epoch 8/8\n",
      "70249/70249 [==============================] - 25s 362us/sample - loss: 496.5521 - val_loss: 454.7661\n"
     ]
    }
   ],
   "source": [
    "model_fc = tf.keras.models.Sequential()\n",
    "model_fc.add(LSTM(64, return_sequences=True, input_shape=(seq_length, 8)))\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(LSTM(64))\n",
    "model_fc.add(BatchNormalization())\n",
    "model_fc.add(Dense(100, input_dim=64))\n",
    "model_fc.add(Activation('relu'))\n",
    "model_fc.add(BatchNormalization())\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(Dense(64, input_dim=64))\n",
    "model_fc.add(Activation('relu'))\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(Dense(18, input_dim=64))\n",
    "\n",
    "model_fc.compile(optimizer='Adam', loss='mse')\n",
    "\n",
    "history = model_fc.fit(features, labels_tip, batch_size=2048, epochs=8, verbose=1, validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 32, 64)            18688     \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 32, 64)            0         \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 64)                33024     \n",
      "_________________________________________________________________\n",
      "batch_normalization (BatchNo (None, 64)                256       \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 100)               6500      \n",
      "_________________________________________________________________\n",
      "activation (Activation)      (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 100)               400       \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 64)                6464      \n",
      "_________________________________________________________________\n",
      "activation_1 (Activation)    (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 18)                1170      \n",
      "=================================================================\n",
      "Total params: 66,502\n",
      "Trainable params: 66,174\n",
      "Non-trainable params: 328\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model_fc.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visual Model Error\n",
    "A similar visualizion is now done on this new more restricted model. These errors show 20% - 25% improvement on fingertip y position."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x14114b908>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.legend(['Train', 'Test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model_fc.predict(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "error = labels_tip - preds\n",
    "sq_error = error * error\n",
    "avg_error = np.mean(sq_error, axis=0)\n",
    "plt.figure(figsize=(15, 10))\n",
    "plt.xticks(rotation=90)\n",
    "plt.ylabel('Prediction Error (mm)')\n",
    "bar = plt.bar(df.columns[8:][[0,1,2,12,13,14,24,25,26,36,37,38,48,49,50,60,61,62]], avg_error)\n",
    "for i in range(0,18,3):\n",
    "    bar[i].set_color('coral')\n",
    "    bar[i+1].set_color('olivedrab')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Large FC Model Architecture\n",
    "Here a much larger model architecture is implemented to understand if this will improve results on the larger dataset. This dataset is approx 650MB, thus a larger model with proportionally higher parameters is expected to be required."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_fc = tf.keras.models.Sequential()\n",
    "model_fc.add(LSTM(256, return_sequences=True, input_shape=(seq_length, 8)))\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(LSTM(256, return_sequences=True))\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(LSTM(128))\n",
    "model_fc.add(BatchNormalization())\n",
    "model_fc.add(Dense(512, input_dim=128))\n",
    "model_fc.add(Activation('relu'))\n",
    "model_fc.add(BatchNormalization())\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(Dense(512, input_dim=512))\n",
    "model_fc.add(Activation('relu'))\n",
    "model_fc.add(BatchNormalization())\n",
    "model_fc.add(Dropout(0.5))\n",
    "model_fc.add(Dense(256, input_dim=512))\n",
    "model_fc.add(Activation('relu'))\n",
    "model_fc.add(Dropout(0.3))\n",
    "model_fc.add(Dense(18, input_dim=64))\n",
    "    \n",
    "model_fc.compile(optimizer='Adam', loss='mse')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_2 (LSTM)                (None, 32, 256)           271360    \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 32, 256)           0         \n",
      "_________________________________________________________________\n",
      "lstm_3 (LSTM)                (None, 32, 256)           525312    \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 32, 256)           0         \n",
      "_________________________________________________________________\n",
      "lstm_4 (LSTM)                (None, 128)               197120    \n",
      "_________________________________________________________________\n",
      "batch_normalization_2 (Batch (None, 128)               512       \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 512)               66048     \n",
      "_________________________________________________________________\n",
      "activation_2 (Activation)    (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_3 (Batch (None, 512)               2048      \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 512)               262656    \n",
      "_________________________________________________________________\n",
      "activation_3 (Activation)    (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "batch_normalization_4 (Batch (None, 512)               2048      \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "activation_4 (Activation)    (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dropout_7 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 18)                4626      \n",
      "=================================================================\n",
      "Total params: 1,463,058\n",
      "Trainable params: 1,460,754\n",
      "Non-trainable params: 2,304\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model_fc.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 70249 samples, validate on 17563 samples\n",
      "Epoch 1/8\n",
      "70249/70249 [==============================] - 210s 3ms/sample - loss: 971.7017 - val_loss: 4776.3685\n",
      "Epoch 2/8\n",
      "70249/70249 [==============================] - 201s 3ms/sample - loss: 353.1096 - val_loss: 323.1201\n",
      "Epoch 3/8\n",
      "70249/70249 [==============================] - 199s 3ms/sample - loss: 220.8060 - val_loss: 263.1873\n",
      "Epoch 4/8\n",
      "70249/70249 [==============================] - 202s 3ms/sample - loss: 187.6194 - val_loss: 419.5479\n",
      "Epoch 5/8\n",
      "70249/70249 [==============================] - 204s 3ms/sample - loss: 169.7060 - val_loss: 264.3906\n",
      "Epoch 6/8\n",
      "70249/70249 [==============================] - 202s 3ms/sample - loss: 165.9210 - val_loss: 321.1586\n",
      "Epoch 7/8\n",
      "70249/70249 [==============================] - 206s 3ms/sample - loss: 149.1318 - val_loss: 229.2482\n",
      "Epoch 8/8\n",
      "70249/70249 [==============================] - 210s 3ms/sample - loss: 170.9081 - val_loss: 298.8186\n"
     ]
    }
   ],
   "source": [
    "history = model_fc.fit(features, labels_tip, batch_size=512, epochs=8, verbose=1, validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x143345d68>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.legend(['Train', 'Test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model_fc.predict(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "error = labels_tip - preds\n",
    "sq_error = error * error\n",
    "avg_error = np.mean(sq_error, axis=0)\n",
    "plt.figure(figsize=(15, 10))\n",
    "plt.xticks(rotation=90)\n",
    "plt.ylabel('Prediction Error (mm)')\n",
    "bar = plt.bar(df.columns[8:][[0,1,2,12,13,14,24,25,26,36,37,38,48,49,50,60,61,62]], avg_error)\n",
    "for i in range(0,18,3):\n",
    "    bar[i].set_color('coral')\n",
    "    bar[i+1].set_color('olivedrab')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}