DataScienceCourse/TFintro.py at master · AngelikaH123/DataScienceCourse · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
#!/usr/bin/env python
# coding: utf-8

# In[1]:


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import tensorflow as tf


# In[3]:


data_url = "https://download.mlcc.google.com/mledu-datasets/california_housing_train.csv"


# In[4]:


df = pd.read_csv(data_url, sep=",")
df.head()


# In[5]:


x = df.households.values.reshape(-1, 1)
y = df.total_rooms.values.reshape(-1,1)


# In[6]:


plt.figure(1, figsize=(8, 6))
plt.scatter(x[::10], y[::10])
plt.xlabel('Number of Households')
plt.ylabel('Total Rooms');
plt.show()


# $$h(x) = x w + b$$

# In[7]:


def get_params(shape):
    np.random.seed(7)
    params = {
    'W':tf.Variable(np.random.randn(*shape), dtype = tf.float32),
    'b':tf.Variable(np.zeros((1, shape[1])), dtype = tf.float32)
    }
    return params


# In[8]:


params = get_params([x.shape[1], y.shape[1]])


# In[9]:


with tf.Session() as sess:
    init = tf.global_variables_initializer()
    sess.run(init)
    result = sess.run(params)
    print(result)


# In[10]:


def linear(x, params):
    h = tf.add(tf.matmul(x, params['W']),params['b'])
    return h


# In[11]:


def get_cost(x, y, model, params, lambd = 0):
    h= model(x, params)
    J = tf.math.reduce_mean((h-y)**2) + ((lambd/2)*tf.math.reduce_sum(params['W']**2))
    return J


# In[12]:


def train(inp, out, model, epochs = 100, lr= 1e-3, lambd = 0):
    x = tf.placeholder(tf.float32, shape = (None, inp.shape[1]))
    y = tf.placeholder(tf.float32, shape = (None, out.shape[1]))
    params = get_params([x.shape[1], y.shape[1]])
    costs = []
    cost = get_cost(x, y, model, params, lambd)
    optimizer = tf.train.GradientDescentOptimizer(learning_rate = lr).minimize(cost)
    init=tf.global_variables_initializer()
    with tf.Session() as sess:
        sess.run(init)
        for epoch in range(epochs):
            c, _ = sess.run([cost, optimizer], feed_dict={x: inp, y:out})

            if epoch%(epochs//10) == 0:
                print('Cost at epoch ' + str(epoch), c)
            if epoch%5 == 0:
                costs.append(c)

        params = sess.run(params)

    return costs, params


# In[ ]:


costs, params = train(x, y, model = linear, lr= 1e-6)


# In[ ]:

plt.figure(2, figsize = (8, 6))
plt.plot(costs)
plt.show()


# In[ ]:


with tf.Session() as sess:
    inp = tf.constant(x, dtype = tf.float32)
    prediction = sess.run(linear(inp, params))


# In[ ]:


plt.figure(3, figsize = (8, 6))
plt.scatter(x[::10], y[::10], label = 'Actual Data')
plt.scatter(x[::10], prediction[::10], label = 'Predicted Data')
plt.legend()
plt.xlabel('Number of Households')
plt.ylabel('Total Rooms');
plt.show()


# In[ ]: