TensorFlow深度学习作业二--癌症转移

下表是肾癌标本资料（数据来源于《卫生统计学》第四版第11章）：

序号	X1	X2	X3	X4	X5	y
1	59	2	43.4	2	1	0
2	36	1	57.2	1	1	0
3	61	2	190	2	1	0
4	58	3	128	4	3	1
5	55	3	80	4	3	1
6	61	1	94.4	4	2	0
7	38	1	76	1	1	0
8	42	1	240	3	2	0
9	50	1	74	1	1	0
10	58	3	68.6	2	2	0
11	68	3	132.8	4	2	0
12	25	2	94.6	4	3	1
13	52	1	56	1	1	0
14	32	1	47.8	2	1	0
15	36	3	31.6	3	1	1
16	42	1	66.2	2	1	0
17	14	3	138.6	3	3	1
18	32	1	114	2	3	0
19	35	1	40.2	2	1	0
20	70	3	177.2	4	3	1
21	65	2	51.6	4	4	1
22	45	2	124	2	4	0
23	68	3	127.2	3	3	1
24	31	2	124.8	2	3	0

数据说明：

y: 肾细胞癌转移情况（有转移y=1；无转移y=0）；

x1: 确诊时患者的年龄（岁）

x2：肾细胞癌血管内皮生长因子（VEGF），其阳性表述由低到高共3个等级；

x3: 肾细胞癌组织内微血管数（MVC）；

x4：肾癌细胞核组织学分级，由低到高共4级；

x5：肾细胞癌分期，由低到高共4期。

请用逻辑回归模型预测肾癌是否转移，自己造至少两个数据进行预测判断。

 

import tensorflow as tf
import numpy as np
import pylab as plt

x_data = np.float32(np.array([[59,2,43.4,2,1],[36,1,57.2,1,1],[61,2,190,2,1],[58,3,128,4,3],[55,3,80,4,3],[61,1,94.4,4,2],[38,1,76,1,1],
                              [42,1,240,3,2],[50,1,74,1,1],[58,3,68.6,2,2],[68,3,132.8,4,2],[25,2,94.6,4,3],[52,1,56,1,1],
                              [32,1,47.8,2,1],[36,3,31.6,3,1],[42,1,66.2,2,1],[14,3,138.6,3,3],[32,1,114,2,3],[35,1,40.2,2,1],
                              [70,3,177.2,4,3],[65,2,51.6,4,4],[45,2,124,2,4],[68,3,127.2,3,3],[31,2,124.8,2,3]]))
y_data = np.float32(np.array([[0,0,0,1,1,0,0,0,0,0,0,1,0,0,1,0,1,0,0,1,1,0,1,0]])).T
#测试数据
yu_x_data = np.float32(np.array([[58,3,128,4,3],[55,3,80,4,3],[36,1,57.2,1,1]]))
yu_b1 = tf.Variable(tf.zeros([3,1]))
#print("x_data:  ", x_data)
#print("y_data:  ", y_data)

w = tf.Variable(tf.zeros([5,1]))
b = tf.Variable(tf.zeros([24,1]))

y1 = tf.matmul(x_data,w) + b
yu_y1 = tf.matmul(yu_x_data,w) + yu_b1
#把值转换到0-1区间
y = tf.nn.sigmoid(y1)
yu_y = tf.nn.sigmoid(yu_y1)
#交叉熵的总值
loss = -tf.reduce_sum(y_data * tf.log(y) + (1 - y_data) * tf.log(1 - y))
train = tf.train.AdamOptimizer(0.02).minimize(loss)
rs = tf.cast(y > 0.9,dtype = tf.int8)
yu_rs = tf.cast(yu_y > 0.9,dtype = tf.int8)
with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    for step in range(10000):
        sess.run(train)
        if step % 1000 == 0:
            print('loss:',sess.run([loss]))
    #print('w:',sess.run(w))
    #print('b:',sess.run(b))
    print('测试数据值的输出：',sess.run(yu_rs))

最终输出：
loss: [27.408615]
loss: [0.10465329]
loss: [0.03114378]
loss: [0.013801137]
loss: [0.0071040075]
loss: [0.0039306525]
loss: [0.0022601355]
loss: [0.0013286448]
loss: [0.0007907951]
loss: [0.00047440067]
要预测的值： [[1]
 [1]
 [0]]