knn_example.py

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Thu Aug 17 10:02:54 2017

@author: saliksyed
"""

from sklearn import datasets
import numpy as np

iris = datasets.load_iris()
features = iris.data[:, :]
labels = iris.target


def dist(pt1, pt2):
    return np.linalg.norm(pt1 - pt2)


def classify_point(query_point, training_data, k=3):
    # compute the distance between query and example
    dists = []
    for point in training_data:
        dists.append(dist(point, query_point))
    
    labels_with_distances = zip(dists, labels)
    k_neighbors = sorted(labels_with_distances, key=lambda x : x[0])[:k]
    
    count = {}
    for neighbor in k_neighbors:
        if not neighbor[1] in count:
            count[neighbor[1]] = 0            
        count[neighbor[1]] += 1
    
    # pick the label that has the highest count!
    # key = the label
    # value = the count
    sorted_counts = sorted(zip(count.keys(), count.values()), key=lambda x : x[1])
    return sorted_counts[0][0]
    

print classify_point(np.array([ 6.5, 3., 5.2, 2. ]), features)