python实现简单kNN

注释写得很清楚了，熟悉了一下python的一些基本语法和numpy中的一些操作。

from numpy import *
import operator

def createDataSet():
    # generate the samples and labels.
    group = array([[1.0,1.1], [1.0,1.0], [0,0], [0,0.1]])
    labels = ['A', 'A', 'B', 'B']
    print group
    return group, labels

def classify(inX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]                  # get the size of one dimension.
    　　　　　　　　　　　　　　　　　　　　　　　　　　　　# calculate the distance between inX and samples.
    diffMat = tile(inX, (dataSetSize, 1)) - dataSet # repeat inX to generate a dataSetSize * 1 matrix. Then subtract the corresponding number in dataSet.
    sqDiffMat = diffMat ** 2                        # get the square of each D-value.
    sqDistances = sqDiffMat.sum(axis=1)             # get the sum of each pair of numbers.
    distances = sqDistances ** 0.5                  # get the square root of each sum. Those are distances between inX and samples.

    sortedDistIndicies = distances.argsort()        # return the index if 'distances' is sorted.
    classCount = {}                                 # make a directory {label:display times}.
    for i in range(k):                                                  # get first kth nearest samples.
        voteIlabel = labels[sortedDistIndicies[i]]                      # get the ith's label.
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1      # count the number of this label.
    sortedClassCount = sorted(classCount.iteritems(),                   # get the most frequent label.
                              key=operator.itemgetter(1), reverse=True)
    return sortedClassCount[0][0]                                       # return the most frequent label.

dataSet, labels = createDataSet()
print classify([-100.0,-100.1], dataSet, labels, 1)