KNN算法：欲分类的点周围找到K个与之最近的点，统计出K个点中最多出现的类别，就是这个点的类别

tensorflow这里是K为1的情况，准确率大约为91%

# -*- coding:utf-8 -*-
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/home/baobao/PycharmProjects/tensorflow_learning/data", one_hot=True)


Xtr, Ytr = mnist.train.next_batch(5000)  # 训练
Xte, Yte = mnist.train.next_batch(1000)  # 测试
x = tf.placeholder(tf.float32, [784])
y = tf.placeholder(tf.float32, [10])  # 预测的值
y_ = tf.placeholder(tf.float32, [10])  # 真实值
distance = tf.reduce_sum(tf.abs(tf.add(Xtr, tf.negative(x))), reduction_indices=1)
index_y = tf.argmin(distance)  # 这个是我们找到的下标，这个是k=1的情况
result = tf.equal(tf.argmax(y), tf.argmax(y_))
accuracy = 0.
init = tf.global_variables_initializer()
with tf.Session() as sess:
    sess.run(init)
    for index, data in enumerate(Xte):
        pre_index = sess.run(index_y, feed_dict={x: data})
        if sess.run(result, feed_dict={y: Ytr[pre_index], y_: Yte[index]}):
            accuracy += 1
    print(accuracy/len(Xte))