1/23/24, 11:11 AM linear_regression.

ipynb - Colaboratory

1 import os
2 import numpy as np
3 import pandas as pd
4 import matplotlib.pyplot as plt
5 # if tensorflow is not installed uncomment the code below
6 #pip install tensorflow
7
8 #import tensorflow as tf # please use this if not successful then use the next two line
9
10 import tensorflow.compat.v1 as tf
11 tf.disable_v2_behavior()

WARNING:tensorflow:From /usr/local/lib/python3.10/dist-packages/tensorflow/python/compat/v2_compat.py:108: disable_resource_variables (from tensorflow.python.ops.variable_scope) i

Instructions for updating:
non-resource variables are not supported in the long term

1 from google.colab import drive

2 drive.mount('/content/drive')

Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).

1 #pip install tensorflow #pip3 install tensorflow

1 path = '/content/drive/MyDrive/Machine-Learning/linear_data/population_profit.txt'
2 data = pd.read_csv(path, header=None, names=['Population','Profit'])

1 data.head()

Population Profit

0 6.1101 17.5920

1 5.5277 9.1302

2 8.5186 13.6620

3 7.0032 11.8540

4 5.8598 6.8233

1 data.tail(4)

Population Profit

93 5.3054 1.98690

94 8.2934 0.14454

95 13.3940 9.05510

96 5.4369 0.61705

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 1/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory
1 data.shape
2 #{}, [], ()

(97, 2)

1 data.Population

0 6.1101
1 5.5277
2 8.5186
3 7.0032
4 5.8598
...
92 5.8707
93 5.3054
94 8.2934
95 13.3940
96 5.4369
Name: Population, Length: 97, dtype: float64

1 data['Profit'].head()

0 17.5920
1 9.1302
2 13.6620
3 11.8540
4 6.8233
Name: Profit, dtype: float64

1 data.describe()

Population Profit

count 97.000000 97.000000

mean 8.159800 5.839135

std 3.869884 5.510262

min 5.026900 -2.680700

25% 5.707700 1.986900

50% 6.589400 4.562300

75% 8.578100 7.046700

max 22.203000 24.147000

1 data.info()
2 #data.describe()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 97 entries, 0 to 96
Data columns (total 2 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Population 97 non-null float64

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 2/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory
1 Profit 97 non-null float64
dtypes: float64(2)
memory usage: 1.6 KB

1 num_of_sample = data.shape[0]
2 num_of_attribute=data.shape[1]
3 print('Number of sample', num_of_sample)
4 print('Number of attribute', num_of_attribute)

Number of sample 97
Number of attribute 2

1 data_x, data_y = data['Population'], data['Profit']

2 data.plot(kind='scatter', x='Population', y='Profit', figsize=(8,6))

<Axes: xlabel='Population', ylabel='Profit'>

1 plt.plot(data_x, data_y,'*')
2 plt.xlabel("Population in 10K's")
3 plt.ylabel("Profit in 10K's")
4 plt.legend(['Data Point'], bbox_to_anchor=(1,1), loc=3)

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 3/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory

<matplotlib.legend.Legend at 0x79e26e128a00>

1 print(data[0:10].shape, data_y[5:9].shape)

(10, 2) (4,)

1 print(data_x[10:15], data_y[8:13])

10 5.7107
11 14.1640
12 5.7340
13 8.4084
14 5.6407
Name: Population, dtype: float64 8 6.5987
9 3.8166
10 3.2522
11 15.5050
12 3.1551
Name: Profit, dtype: float64

1 #Simple Linear regression y=mx+b Creating X as Population and Y as Profits

2 with tf.name_scope('inputs'):
3 X = tf.placeholder(tf.float32, name="input")
4 Y = tf.placeholder(tf.float32, name="output")

1 # Creating weight and bias variables initialized to Zero

2 with tf.name_scope('parameters'):
3 w = tf.Variable(0.0, name='Weights')
4 b = tf.Variable(0.0, name='bias')

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 4/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory
1 # Defining a linear regression model to predict profit output from population input
2 # Y_predict = X*w + b
3 with tf.name_scope('regression_model'):
4 Y_predict = X*w + b

1 # Defining loss
2 with tf.name_scope('loss_function'):
3 loss = tf.reduce_mean(tf.square(Y-Y_predict, name = 'loss'))

1 #### Gradient descent optimizer for optimizing weight and bias

2 optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.001)
3 train_op = optimizer.minimize(loss)

1 #### Creating a summary

2 w_hist = tf.summary.histogram("Weights", w)
3 b_hist = tf.summary.histogram("biasBiases", b)
4 y_hist = tf.summary.histogram("y_predicted", Y_predict)
5 cost = tf.summary.scalar("loss", loss)
6 merged_summary = tf.summary.merge_all()

1 #### Saving the trained model

2 saver = tf.train.Saver()

1 #### Calling TF operation through session

2 cost_history = np.empty(shape=[1], dtype=float)
3 with tf.Session() as sess:
4 ### Summary writer
5 Summary_writter = tf.summary.FileWriter('/content/drive/MyDrive/Machine-Learning/linear_data/linear-model-test/linear_reg_summary', sess.graph)
6 sess.run(tf.global_variables_initializer())
7 for i in range(800):
8 for x, y in zip(data_x, data_y):
9 _, loss_v, summary = sess.run([train_op, loss, merged_summary], feed_dict={X:x, Y:y})
10 cost_history = np.append(cost_history, loss_v)
11 if i % 20 == 0:
12 print("the loss is : ", loss_v, " at ", i)
13 Summary_writter.add_summary(summary,i)
14 w_value, b_value = sess.run([w,b])
15 saver.save(sess,"/content/drive/MyDrive/Machine-Learning/linear_data/linear-model-test/model-final-lin-reg")
16

the loss is : 5.8409615 at 0

the loss is : 3.011311 at 20
the loss is : 1.8686026 at 40
the loss is : 1.3639505 at 60
the loss is : 1.1241888 at 80
the loss is : 1.0042403 at 100
the loss is : 0.94222325 at 120
the loss is : 0.9095278 at 140
the loss is : 0.89209586 at 160
the loss is : 0.88274276 at 180
the loss is : 0.8777063 at 200
the loss is : 0.87499225 at 220
the loss is : 0.8735263 at 240
the loss is : 0.87273496 at 260

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 5/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory
the loss is : 0.8723056 at 280
the loss is : 0.8720758 at 300
the loss is : 0.8719516 at 320
the loss is : 0.87187946 at 340
the loss is : 0.87183625 at 360
the loss is : 0.8718193 at 380
the loss is : 0.8718144 at 400
the loss is : 0.87181044 at 420
the loss is : 0.87180996 at 440
the loss is : 0.87180996 at 460
the loss is : 0.87180996 at 480
the loss is : 0.87180996 at 500
the loss is : 0.87180996 at 520
the loss is : 0.87180996 at 540
the loss is : 0.87180996 at 560
the loss is : 0.87180996 at 580
the loss is : 0.87180996 at 600
the loss is : 0.87180996 at 620
the loss is : 0.87180996 at 640
the loss is : 0.87180996 at 660
the loss is : 0.87180996 at 680
the loss is : 0.87180996 at 700
the loss is : 0.87180996 at 720
the loss is : 0.87180996 at 740
the loss is : 0.87180996 at 760
the loss is : 0.87180996 at 780

1 w_value, b_value

(0.9960518, -3.9217443)

1 len= cost_history.shape[0]
2 iters = len
3 fig, ax = plt.subplots(figsize=(6,4))
4 ax.plot(np.arange(iters), cost_history, 'r')
5 ax.set_xlabel('Iterations')
6 ax.set_ylabel('Cost')
7 ax.set_title('Error vs Training Epoch')

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 6/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory

Text(0.5, 1.0, 'Error vs Training Epoch')

1 w_value

0.9960518

1 b_value

-3.9217443

1 print(data_x[0:10],data_y[0:10])

0 6.1101
1 5.5277
2 8.5186
3 7.0032
4 5.8598
5 8.3829
6 7.4764
7 8.5781
8 6.4862
9 5.0546
Name: Population, dtype: float64 0 17.5920
1 9.1302
2 13.6620
3 11.8540
4 6.8233
5 11.8860
6 4.3483
7 12.0000
8 6.5987
9 3.8166
Name: Profit, dtype: float64

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 7/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory
1 data_x[0:10]

0 6.1101
1 5.5277
2 8.5186
3 7.0032
4 5.8598
5 8.3829
6 7.4764
7 8.5781
8 6.4862
9 5.0546
Name: Population, dtype: float64

1 np.array(data_x[0:10])

array([6.1101, 5.5277, 8.5186, 7.0032, 5.8598, 8.3829, 7.4764, 8.5781,

6.4862, 5.0546])

1 print(np.array(data_x[0:10]))
2 print(np.array(data_y[0:10]))

[6.1101 5.5277 8.5186 7.0032 5.8598 8.3829 7.4764 8.5781 6.4862 5.0546]
[17.592 9.1302 13.662 11.854 6.8233 11.886 4.3483 12. 6.5987
3.8166]

1 ####### test
2 x_test = np.array(data_x[0:10])
3 y_test = np.array(data_y[0:10])
4 y_test_predicted = x_test*w_value + b_value
5 print('predicted value',y_test_predicted)
6 print('TRUE VALUE',y_test)
7 print(w_value, b_value)

predicted value [2.16423169 1.58413112 4.56322242 3.05380554 1.91491992 4.42805819

3.52513724 4.6224875 2.53884676 1.11289902]
TRUE VALUE [17.592 9.1302 13.662 11.854 6.8233 11.886 4.3483 12. 6.5987
3.8166]
0.9960518 -3.9217443

1 print(w_value, b_value) ###### test

2 x_test = np.array(data_x)
3 y_test = np.array(data_y)
4 y_test_predicted = x_test*w_value + b_value
5 ###print(y_test_predicted)

0.9960518 -3.9217443

1 plt.plot(x_test, y_test, "o", x_test, y_test_predicted, "-")

2 plt.xlabel("Population in 10K's")
3 plt.ylabel("Profit in 10K USD")
4 plt.legend(['Data', 'Linear Regression'], bbox_to_anchor = (1,1), loc = 4)
5 plt.show()

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 8/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory

1 plt.scatter(x_test, y_test,marker= "o")

2 plt.plot(x_test, y_test_predicted,"b")
3 plt.xlabel("Population in 10K's")
4 plt.ylabel("Profit in 10K USD")
5 plt.legend(['Data', 'Linear Regression'], bbox_to_anchor = (1,1), loc = 4)
6 plt.show()

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 9/10
1/23/24, 11:11 AM linear_regression.ipynb - Colaboratory

1 tf.reset_default_graph()
2 with tf.Session() as sess:
3 imp meta=tf.train.import meta graph("/content/drive/MyDrive/Machine-Learning/linear data/linear-model-test/model-final-lin-reg.meta")

https://colab.research.google.com/drive/19QQe4ulOcsyQS5d6irtediXca_rmUhkp?authuser=1#scrollTo=FBaM7O0YLlCI&printMode=true 10/10