KNN实现手写数字分类---Kaggle

#!/usr/bin/python
# coding=utf-8
from numpy import *
import operator
import pandas as pd
import  sys
import  KNN
sys.path.append(r'C:/Users/Documents/6、play/')
import xlwt


##导入训练集：
def load_data():

    label=pd.read_csv('label.csv', header=None)
    label=label.values
    data=[]
    for i in label:
        for j in i :
            data.append(j)
    label=data
    dataset=pd.read_csv('dataset.csv', header=None)
    dataset=dataset.values
    test = pd.read_csv('test.csv', header=None)
    test = test.values
    return label,dataset,test


def kNNClassify(newInput, dataSet, labels, k):
    numSamples = dataSet.shape[0]  # shape[0]表示行数
    diff = tile(newInput, (numSamples, 1)) - dataSet  # 按元素求差值

    squaredDiff = diff ** 2  # 将差值平方

    squaredDist = sum(squaredDiff, axis=1)  # 按行累加

    distance = squaredDist ** 0.5  # 将差值平方和求开方，即得距离
    sortedDistIndices = argsort(distance)
    classCount = {}  # define a dictionary (can be append element)
    for i in range(0,k):
        voteLabel = labels[sortedDistIndices[i]]
        classCount[voteLabel] = classCount.get(voteLabel, 0) + 1
  # # step 5: 返回出现次数最多的类别标签
    maxCount = 0
    for key, value in classCount.items():
        if value > maxCount:
            maxCount = value
    maxIndex = key
    return maxIndex


book=xlwt.Workbook(encoding='utf-8',style_compression=0)
sheet=book.add_sheet('test',cell_overwrite_ok=True)


label,dataset,test = load_data()
n=1
for i in test:
    outputLabel = kNNClassify(i, dataset, label, 3)
    print('第 ',n,'个为：  ',str(outputLabel))
    sheet.write(0, 0, str(outputLabel))
    n=n+1
book.save('out.xls')