doco

timm · Aug 2, 2024 · 4aa6e8a · 4aa6e8a
1 parent 481c67a
commit 4aa6e8a
Show file tree

Hide file tree

Showing 2 changed files with 64 additions and 0 deletions.
diff --git a/etc/header.sh b/etc/header.sh
@@ -0,0 +1,18 @@
+cat <<EOF
+<!DOCTYPE html>
+<html lang="en">
+<head>
+<title>$1</title>
+<meta charset="UTF-8">
+<link rel="icon" type="image/x-icon" href="favicon.ico">
+<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.9.0/styles/vs.min.css">
+<script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.9.0/highlight.min.js"></script>
+<script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.9.0/languages/python.min.js"></script>
+<script>hljs.highlightAll();</script>
+<link rel="stylesheet" href="ezr.css">
+</head><body>
+<small><p align="right"><a href="home">home</a> :: <a href="issues">issues</a> :: <a href="license">license</a>
+:: <a href="home">home</a> :: <a href="issues">issues</a> :: <a href="license">license</a></p></small>
+
+<h1>$1</h1>
+EOF
diff --git a/src/knn.py b/src/knn.py
@@ -0,0 +1,46 @@
+# # KNN
+# Here is some code that does a 5x5 cross-val for a knn.
+# Adapt it to  explore k=1,2,5 neasest neghits.
+# also looking at the code HERE, try exploring a random subset of 
+# `train` of size 25,50,200,400
+# 
+# ```python
+import random,sys
+sys.path.insert(0, '..')
+from  stats import SOME,some0
+from  ezr import DATA, SYM, csv, xval, the
+
+def dot(): print(".", file=sys.stderr, flush=True, end="")
+
+def knn(data,  k, row1, rows=None):
+  seen = SYM()
+  for row in data.neighbors(row1, rows=rows)[:k]:
+    seen.add( row[data.cols.klass.at] )
+  return seen.mode
+
+def one(data,k,p, train,test):
+  n,acc = 0,0
+  the.p=p
+  for row in test:
+    want = row[ data.cols.klass.at ]
+    got  = knn(data, k, row, train)
+    acc += want==got
+    n   += 1
+  return acc/n
+
+def main(file): 
+  random.seed(1234567891)
+  data = DATA().adds(csv(file))
+  somes = []
+  for n in [25,50,100,200,100000]:
+    for k in [1,2,5]:
+      for p in [1,2,3,4]:
+        dot()
+        somes  += [SOME(txt=f"k{k}p{p}n{n}")]
+        for train,test in xval(data.rows, 5,5, n):
+          somes[-1].add(one(data, k, p, train, test))
+  print("\n" + file)
+  some0(somes)
+
+for file in sys.argv[1:]: main(file)
+# ```