kNN —— 手写数字识别

最新推荐文章于 2024-06-20 21:22:09 发布

- birdguan -

最新推荐文章于 2024-06-20 21:22:09 发布

阅读量264

点赞数

分类专栏： Machine Learning

本文链接：https://blog.csdn.net/qq_30241709/article/details/87860032

版权

Machine Learning 专栏收录该内容

31 篇文章 0 订阅

订阅专栏

这里手写数据的数据集并不是大家熟知的MNIST，而是土耳其伊斯坦布尔海峡大学计算机工程系的E.Alpaydin与C.Kaynak的数据集，地址：

http://archive.ics.uci.edu/ml

每一个数字为 $32px \times 32px$ 黑白图像，如下图所示：

程序如下：

import numpy as np
import os
import operator


def img2vector(filename):
    return_vec = np.zeros((1, 1024))
    fr = open(filename)
    for i in range(32):
        line_str = fr.readline()
        for j in range(32):
            return_vec[0, 32 * i + j] = int(line_str[j])
    return return_vec


def classify0(inX, data_set, labels, k):
    data_set_size = data_set.shape[0]
    diff_mat = np.tile(inX, (data_set_size, 1)) - data_set
    sq_diff_mat = diff_mat ** 2
    sq_distances = sq_diff_mat.sum(axis=1)
    distances = sq_distances ** 0.5
    sorted_dist_indicies = distances.argsort()
    class_count = {}
    for i in range(k):
        vote_ilabel = labels[sorted_dist_indicies[i]]
        class_count[vote_ilabel] = class_count.get(vote_ilabel, 0) + 1
    sorted_class_count = sorted(class_count.items(), key=operator.itemgetter(1), reverse=True)
    return sorted_class_count[0][0]


def handwriting_class_test():
    hw_labels = []
    training_file_list = os.listdir('digits/trainingDigits')
    m = len(training_file_list)
    training_mat = np.zeros((m, 1024))
    for i in range(m):
        file_name_string = training_file_list[i]
        file_string = file_name_string.split('.')[0]
        digit_class = int(file_string.split('_')[0])
        hw_labels.append(digit_class)
        training_mat[i, :] = img2vector('digits/trainingDigits/%s' % file_name_string)
    test_file_list = os.listdir('digits/testDigits')
    error_count = 0.0
    error_rate = 0.0
    num_test = len(test_file_list)
    for i in range(num_test):
        file_name_string = test_file_list[i]
        file_string = file_name_string.split('.')[0]
        test_digit_class = int(file_string.split('_')[0])
        test_vector = img2vector('digits/testDigits/%s' % file_name_string)
        classify_result = classify0(test_vector, training_mat, hw_labels, 3)
        if classify_result != test_digit_class:
            error_count += 1
        error_rate = error_count/(i+1)
        print("classify result: %d, real result: %d, current error rate: %f%%" % (classify_result, test_digit_class, error_rate*100))
    print("=====> end of the program <=====")
    print("total error rate: %f%%" % (error_rate*100))


if __name__ == '__main__':
    handwriting_class_test()

结果如下：

classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%
classify result: 0, real result: 0, current error rate: 0.000000%

...

classify result: 9, real result: 9, current error rate: 1.084599%
classify result: 9, real result: 9, current error rate: 1.083424%
classify result: 9, real result: 9, current error rate: 1.082251%
classify result: 9, real result: 9, current error rate: 1.081081%
classify result: 9, real result: 9, current error rate: 1.079914%
classify result: 9, real result: 9, current error rate: 1.078749%
classify result: 9, real result: 9, current error rate: 1.077586%
classify result: 9, real result: 9, current error rate: 1.076426%
classify result: 9, real result: 9, current error rate: 1.075269%
classify result: 9, real result: 9, current error rate: 1.074114%
classify result: 9, real result: 9, current error rate: 1.072961%
classify result: 9, real result: 9, current error rate: 1.071811%
classify result: 9, real result: 9, current error rate: 1.070664%
classify result: 9, real result: 9, current error rate: 1.069519%
classify result: 9, real result: 9, current error rate: 1.068376%
classify result: 9, real result: 9, current error rate: 1.067236%
classify result: 9, real result: 9, current error rate: 1.066098%
classify result: 9, real result: 9, current error rate: 1.064963%
classify result: 9, real result: 9, current error rate: 1.063830%
classify result: 9, real result: 9, current error rate: 1.062699%
classify result: 9, real result: 9, current error rate: 1.061571%
classify result: 9, real result: 9, current error rate: 1.060445%
classify result: 9, real result: 9, current error rate: 1.059322%
classify result: 9, real result: 9, current error rate: 1.058201%
classify result: 9, real result: 9, current error rate: 1.057082%
=====> end of the program <=====
total error rate: 1.057082%