model_D.ipynb

{"cells":[{"cell_type":"code","execution_count":2,"metadata":{},"outputs":[],"source":["import numpy as np\n","from pandas import DataFrame\n","\n","\n","inp = np.load(\"data/original/dataset_smile_challenge.npy\", allow_pickle=True).item()\n","train = inp['train']\n","test = inp['test']\n","threshold = 1 # no stress below th\n","test_label = [0 if i < threshold else 1 for i in test['labels']]\n","train_label = [0 if i < threshold else 1 for i in train['labels']]\n"]},{"cell_type":"markdown","metadata":{},"source":["## Masking"]},{"cell_type":"code","execution_count":3,"metadata":{},"outputs":[],"source":["data_hand_features = train['hand_crafted_features']\n","maskECG = data_hand_features['ECG_masking']\n","maskGSR = data_hand_features['GSR_masking']\n","mask = np.any(maskECG*maskGSR == 1, axis=1)\n","data_ecg = data_hand_features['ECG_features']\n","data_gsr = data_hand_features['GSR_features']\n","data_ecg = np.array(data_ecg)[mask]\n","data_gsr = np.array(data_gsr)[mask]\n","data_hand_features['ECG_features'] = data_ecg\n","data_hand_features['GSR_features'] = data_gsr\n","train['hand_crafted_features'] = data_hand_features\n","\n","data_deep_features = train['deep_features']\n","mask_deep = np.any(data_deep_features['masking'] == 1, axis = 1)\n","dataECG_C = data_deep_features['ECG_features_C']\n","dataECG_T = data_deep_features['ECG_features_T']\n","dataECG_C = dataECG_C[mask_deep]\n","dataECG_T = dataECG_T[mask_deep]\n","data_deep_features['ECG_features_C'] = dataECG_C\n","data_deep_features['ECG_features_T'] = dataECG_T\n","train['deep_features'] = data_deep_features"]},{"cell_type":"code","execution_count":4,"metadata":{},"outputs":[],"source":["def getFeatures(features, rangestart = 0, rangeend = 60):\n","    from sklearn.linear_model import LinearRegression\n","    fi = []\n","    for j in range(0,np.shape(features)[1]):\n","        ft = [x if ~np.isnan(x) else 0 for x in features[rangestart:rangeend,j]]\n","       \n","        m = np.mean(ft)\n","        std = np.std(ft)\n","        diff = np.max(np.abs(np.diff(ft)))\n","        max = np.max(np.abs(ft))\n","        med = np.median(ft)\n","\n","        fi.append(m)\n","        fi.append(std)\n","        fi.append(diff)\n","        fi.append(max)\n","        fi.append(med)\n","    return fi"]},{"cell_type":"code","execution_count":5,"metadata":{},"outputs":[],"source":["def calcFeatures(data):\n","    data_deep_features = data['deep_features']\n","    data_hand_features = data['hand_crafted_features']\n","    data_ecg = data_hand_features['ECG_features']\n","    data_gsr = data_hand_features['GSR_features']\n","    dataECG_C = data_deep_features['ECG_features_C']\n","    dataECG_T = data_deep_features['ECG_features_T']\n","    \n","    l = len(data_ecg)\n","    f = []\n","    for i in range(0,l):\n","        \n","        ecg_c = dataECG_C[i]\n","        ecg_t = dataECG_T[i]\n","        gsr = data_gsr[i]\n","        ecg = data_ecg[i]\n","        \n","        features = gsr\n","        features = np.append(features, ecg, axis=1)\n","    \n","        f.append(getFeatures(features))\n","    return f"]},{"cell_type":"code","execution_count":6,"metadata":{},"outputs":[],"source":["dotrain = True\n","if dotrain:\n","   f = calcFeatures(train)\n"]},{"cell_type":"code","execution_count":7,"metadata":{},"outputs":[],"source":["select = True\n","if select:\n","    from sklearn.feature_selection import SelectKBest\n","    from sklearn.feature_selection import *\n","    # define feature selection\n","    fs = SelectKBest(score_func=chi2, k=15)\n","    # apply feature selection\n","    fs.fit(f, np.array(train_label)[mask])\n","    selected = fs.transform(f)\n","    selected_train_features = selected\n","    mask_features = fs.get_support()\n","    "]},{"cell_type":"code","execution_count":8,"metadata":{},"outputs":[],"source":["dotest = True\n","if dotest:\n","    f = calcFeatures(test)\n","\n","    selected_test_features = np.array(f)[:,mask_features]\n"]},{"cell_type":"markdown","metadata":{},"source":["## Logistic Regression"]},{"cell_type":"code","execution_count":9,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["Accuracy on train: 62.500%\n"]}],"source":["test_label = [0 if i < threshold else 1 for i in inp['test']['labels']]\n","\n","from sklearn.linear_model import LogisticRegression\n","from sklearn.model_selection import *\n","\n","modelML = LogisticRegression()\n","\n","modelML.fit(selected_train_features, np.array(train_label)[mask])\n","result = modelML.score(selected_train_features, np.array(train_label)[mask])\n","print(\"Accuracy on train: %.3f%%\" % (result*100.0))\n"]},{"cell_type":"code","execution_count":11,"metadata":{},"outputs":[{"name":"stdout","output_type":"stream","text":["zip created\n"]}],"source":["import os\n","predicted = modelML.predict(np.array(selected_test_features))\n","np.savetxt(os.path.join('submission','modelD', 'answer.txt'),predicted,fmt='%1i')\n","from zipfile import ZipFile\n","\n","# create a ZipFile object\n","zipObj = ZipFile('submission/modelD/answer.zip', 'w')\n","# Add multiple files to the zip\n","zipObj.write('submission/modelD/answer.txt',arcname='answer.txt')\n","# close the Zip File\n","zipObj.close()\n","\n","print('zip created')"]}],"metadata":{"kernelspec":{"display_name":"Python 3.10.4 ('env': venv)","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":"3.10.4"},"vscode":{"interpreter":{"hash":"02ed20ca32e85c58ac17d6c1a94427bbb9ae10306c3e0caf395fb5182c7ec815"}}},"nbformat":4,"nbformat_minor":2}