import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline

df = pd.read_csv('/content/fake_reg.csv')
df.head()

# consider this as regression problem, where
# based on feature 1 and feature2, we need to predict the price

sns.pairplot(df)

<seaborn.axisgrid.PairGrid at 0x7fa2d26a5ed0>

# create test, train split

from sklearn.model_selection import train_test_split

# and convert the dataset into values because tensorflow dont accept pandas data frame or series
X = df[['feature1', 'feature2']].values
y = df['price'].values

X

array([[ 999.78755752,  999.7660962 ],
       [ 998.86161491, 1001.04240315],
       [1000.07026691,  998.84401463],
       ...,
       [1001.45164617,  998.84760554],
       [1000.77102275,  998.56285086],
       [ 999.2322436 , 1001.45140713]])

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=42)

X_train.shape

(700, 2)

X_test.shape

(300, 2)

# normalise or scale the dataset
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()

# help(MinMaxScaler) 
# if requred, read throug the help section by typing 'help(MinMaxScalar)
# No need to scale the label but the features only

scaler.fit(X_train)

MinMaxScaler()

X_train = scaler.transform(X_train)

X_test = scaler.transform(X_test)

X_train.min()

0.0

X_train.max()

1.0

X_train

array([[0.74046017, 0.32583248],
       [0.43166001, 0.2555088 ],
       [0.18468554, 0.70500664],
       ...,
       [0.54913363, 0.79933822],
       [0.2834197 , 0.38818708],
       [0.56282703, 0.42371827]])

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense

model = Sequential()
model.add(Dense(4, activation='relu'))
model.add(Dense(4, activation = "relu"))
model.add(Dense(4, activation = "relu"))
model.add(Dense(1))
model.compile(optimizer="rmsprop", loss='mse')

model.fit(x=X_train, y=y_train, epochs = 250)

Epoch 1/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2021
Epoch 2/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2882
Epoch 3/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5449
Epoch 4/250
22/22 [==============================] - 0s 2ms/step - loss: 23.4926
Epoch 5/250
22/22 [==============================] - 0s 3ms/step - loss: 24.6201
Epoch 6/250
22/22 [==============================] - 0s 3ms/step - loss: 23.7742
Epoch 7/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3183
Epoch 8/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1381
Epoch 9/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3549
Epoch 10/250
22/22 [==============================] - 0s 4ms/step - loss: 24.3110
Epoch 11/250
22/22 [==============================] - 0s 3ms/step - loss: 24.6376
Epoch 12/250
22/22 [==============================] - 0s 3ms/step - loss: 24.0384
Epoch 13/250
22/22 [==============================] - 0s 3ms/step - loss: 24.1267
Epoch 14/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2055
Epoch 15/250
22/22 [==============================] - 0s 3ms/step - loss: 24.7081
Epoch 16/250
22/22 [==============================] - 0s 3ms/step - loss: 24.0606
Epoch 17/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1664
Epoch 18/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3561
Epoch 19/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3632
Epoch 20/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4066
Epoch 21/250
22/22 [==============================] - 0s 3ms/step - loss: 24.0745
Epoch 22/250
22/22 [==============================] - 0s 3ms/step - loss: 24.1912
Epoch 23/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4251
Epoch 24/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4877
Epoch 25/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2361
Epoch 26/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9286
Epoch 27/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2013
Epoch 28/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4589
Epoch 29/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4260
Epoch 30/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9780
Epoch 31/250
22/22 [==============================] - 0s 4ms/step - loss: 24.1774
Epoch 32/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3303
Epoch 33/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2686
Epoch 34/250
22/22 [==============================] - 0s 3ms/step - loss: 24.5968
Epoch 35/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2579
Epoch 36/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4497
Epoch 37/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3964
Epoch 38/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4294
Epoch 39/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4171
Epoch 40/250
22/22 [==============================] - 0s 3ms/step - loss: 24.5096
Epoch 41/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1096
Epoch 42/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5023
Epoch 43/250
22/22 [==============================] - 0s 3ms/step - loss: 24.7560
Epoch 44/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4277
Epoch 45/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3015
Epoch 46/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2988
Epoch 47/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4406
Epoch 48/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3641
Epoch 49/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3656
Epoch 50/250
22/22 [==============================] - 0s 2ms/step - loss: 24.8987
Epoch 51/250
22/22 [==============================] - 0s 3ms/step - loss: 24.0434
Epoch 52/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3114
Epoch 53/250
22/22 [==============================] - 0s 3ms/step - loss: 24.0254
Epoch 54/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4083
Epoch 55/250
22/22 [==============================] - 0s 3ms/step - loss: 24.4475
Epoch 56/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3145
Epoch 57/250
22/22 [==============================] - 0s 5ms/step - loss: 24.4794
Epoch 58/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3725
Epoch 59/250
22/22 [==============================] - 0s 3ms/step - loss: 24.6375
Epoch 60/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5955
Epoch 61/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4699
Epoch 62/250
22/22 [==============================] - 0s 3ms/step - loss: 23.9523
Epoch 63/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4700
Epoch 64/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3923
Epoch 65/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0723
Epoch 66/250
22/22 [==============================] - 0s 2ms/step - loss: 23.8749
Epoch 67/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4307
Epoch 68/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1323
Epoch 69/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2195
Epoch 70/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2206
Epoch 71/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5046
Epoch 72/250
22/22 [==============================] - 0s 3ms/step - loss: 24.3642
Epoch 73/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2649
Epoch 74/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3417
Epoch 75/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3061
Epoch 76/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3012
Epoch 77/250
22/22 [==============================] - 0s 3ms/step - loss: 24.2340
Epoch 78/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2617
Epoch 79/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4010
Epoch 80/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1290
Epoch 81/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0816
Epoch 82/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2827
Epoch 83/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1365
Epoch 84/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2153
Epoch 85/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2936
Epoch 86/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2648
Epoch 87/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0684
Epoch 88/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5049
Epoch 89/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2773
Epoch 90/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3830
Epoch 91/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2693
Epoch 92/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2628
Epoch 93/250
22/22 [==============================] - 0s 1ms/step - loss: 24.3257
Epoch 94/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2392
Epoch 95/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5691
Epoch 96/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3787
Epoch 97/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2215
Epoch 98/250
22/22 [==============================] - 0s 2ms/step - loss: 23.8202
Epoch 99/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1144
Epoch 100/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2367
Epoch 101/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1516
Epoch 102/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2913
Epoch 103/250
22/22 [==============================] - 0s 2ms/step - loss: 24.7692
Epoch 104/250
22/22 [==============================] - 0s 1ms/step - loss: 24.1539
Epoch 105/250
22/22 [==============================] - 0s 1ms/step - loss: 23.8611
Epoch 106/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9971
Epoch 107/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1035
Epoch 108/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3654
Epoch 109/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0588
Epoch 110/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4332
Epoch 111/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2540
Epoch 112/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2551
Epoch 113/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5121
Epoch 114/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6285
Epoch 115/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5548
Epoch 116/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5045
Epoch 117/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5942
Epoch 118/250
22/22 [==============================] - 0s 2ms/step - loss: 23.6348
Epoch 119/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5779
Epoch 120/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6205
Epoch 121/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4149
Epoch 122/250
22/22 [==============================] - 0s 2ms/step - loss: 23.8001
Epoch 123/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4824
Epoch 124/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1463
Epoch 125/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2978
Epoch 126/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2441
Epoch 127/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3653
Epoch 128/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3846
Epoch 129/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2313
Epoch 130/250
22/22 [==============================] - 0s 1ms/step - loss: 24.5292
Epoch 131/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0321
Epoch 132/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1436
Epoch 133/250
22/22 [==============================] - 0s 1ms/step - loss: 24.3465
Epoch 134/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6819
Epoch 135/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0843
Epoch 136/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1636
Epoch 137/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5991
Epoch 138/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5355
Epoch 139/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2291
Epoch 140/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4753
Epoch 141/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3483
Epoch 142/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1712
Epoch 143/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6259
Epoch 144/250
22/22 [==============================] - 0s 2ms/step - loss: 24.8681
Epoch 145/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3802
Epoch 146/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4746
Epoch 147/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9587
Epoch 148/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3929
Epoch 149/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9384
Epoch 150/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5323
Epoch 151/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4623
Epoch 152/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6169
Epoch 153/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4987
Epoch 154/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1072
Epoch 155/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1062
Epoch 156/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1921
Epoch 157/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0414
Epoch 158/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3024
Epoch 159/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2404
Epoch 160/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3603
Epoch 161/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0775
Epoch 162/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3982
Epoch 163/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4710
Epoch 164/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2896
Epoch 165/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3714
Epoch 166/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2259
Epoch 167/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6570
Epoch 168/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9927
Epoch 169/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6605
Epoch 170/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3853
Epoch 171/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1116
Epoch 172/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1977
Epoch 173/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6523
Epoch 174/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1437
Epoch 175/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5628
Epoch 176/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1733
Epoch 177/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1545
Epoch 178/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6782
Epoch 179/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5265
Epoch 180/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5222
Epoch 181/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2190
Epoch 182/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0897
Epoch 183/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5576
Epoch 184/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5399
Epoch 185/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3545
Epoch 186/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9272
Epoch 187/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1280
Epoch 188/250
22/22 [==============================] - 0s 2ms/step - loss: 24.8770
Epoch 189/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5467
Epoch 190/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3382
Epoch 191/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4816
Epoch 192/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1703
Epoch 193/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5132
Epoch 194/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4400
Epoch 195/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2927
Epoch 196/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1765
Epoch 197/250
22/22 [==============================] - 0s 2ms/step - loss: 24.7389
Epoch 198/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5436
Epoch 199/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4114
Epoch 200/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5023
Epoch 201/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3604
Epoch 202/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1969
Epoch 203/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2007
Epoch 204/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3450
Epoch 205/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6767
Epoch 206/250
22/22 [==============================] - 0s 2ms/step - loss: 23.9474
Epoch 207/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4251
Epoch 208/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2889
Epoch 209/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0246
Epoch 210/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0154
Epoch 211/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5941
Epoch 212/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1882
Epoch 213/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2154
Epoch 214/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3963
Epoch 215/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0128
Epoch 216/250
22/22 [==============================] - 0s 3ms/step - loss: 24.6367
Epoch 217/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2675
Epoch 218/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5879
Epoch 219/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2869
Epoch 220/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2627
Epoch 221/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1074
Epoch 222/250
22/22 [==============================] - 0s 2ms/step - loss: 24.6439
Epoch 223/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4381
Epoch 224/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5369
Epoch 225/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1022
Epoch 226/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2265
Epoch 227/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1496
Epoch 228/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1179
Epoch 229/250
22/22 [==============================] - 0s 2ms/step - loss: 24.5185
Epoch 230/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3337
Epoch 231/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1804
Epoch 232/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2180
Epoch 233/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2324
Epoch 234/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2262
Epoch 235/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0912
Epoch 236/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4821
Epoch 237/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3106
Epoch 238/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4344
Epoch 239/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4802
Epoch 240/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4350
Epoch 241/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1894
Epoch 242/250
22/22 [==============================] - 0s 2ms/step - loss: 24.1196
Epoch 243/250
22/22 [==============================] - 0s 2ms/step - loss: 24.0202
Epoch 244/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3948
Epoch 245/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2129
Epoch 246/250
22/22 [==============================] - 0s 2ms/step - loss: 24.7444
Epoch 247/250
22/22 [==============================] - 0s 2ms/step - loss: 24.3731
Epoch 248/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2083
Epoch 249/250
22/22 [==============================] - 0s 2ms/step - loss: 24.4584
Epoch 250/250
22/22 [==============================] - 0s 2ms/step - loss: 24.2388

<keras.callbacks.History at 0x7fa25bc759d0>

model.history.history

{'loss': [256572.046875,
  256367.515625,
  256177.96875,
  255969.921875,
  255740.828125,
  255486.265625,
  255206.03125,
  254901.25,
  254568.4375,
  254202.859375,
  253800.84375,
  253350.265625,
  252851.265625,
  252299.71875,
  251694.015625,
  251027.15625,
  250294.75,
  249493.078125,
  248619.171875,
  247676.390625,
  246652.25,
  245534.1875,
  244333.1875,
  243036.234375,
  241640.4375,
  240145.84375,
  238531.9375,
  236814.5625,
  234994.671875,
  233047.9375,
  230978.90625,
  228776.609375,
  226436.796875,
  223975.046875,
  221366.1875,
  218615.90625,
  215713.890625,
  212662.5625,
  209457.6875,
  206100.953125,
  202581.140625,
  198914.28125,
  195057.984375,
  191074.40625,
  186926.78125,
  182605.71875,
  178122.71875,
  173493.625,
  168702.921875,
  163747.59375,
  158636.171875,
  153380.859375,
  147997.640625,
  142461.8125,
  136816.359375,
  131068.8359375,
  125187.2421875,
  119193.6796875,
  113158.640625,
  107010.2421875,
  100834.1484375,
  94595.5546875,
  88363.609375,
  82119.484375,
  75879.0859375,
  69665.140625,
  63536.6328125,
  57533.8515625,
  51664.30859375,
  45917.84375,
  40393.5078125,
  35064.0078125,
  30006.33203125,
  25265.31640625,
  20843.11328125,
  16854.552734375,
  13272.578125,
  10164.947265625,
  7588.80712890625,
  5602.05810546875,
  4188.60546875,
  3337.8125,
  2982.46533203125,
  2896.08154296875,
  2854.520263671875,
  2815.706787109375,
  2771.947998046875,
  2732.728515625,
  2689.320068359375,
  2647.334228515625,
  2610.560791015625,
  2574.21923828125,
  2532.7578125,
  2497.965576171875,
  2460.290771484375,
  2417.96435546875,
  2381.38134765625,
  2346.566650390625,
  2312.5732421875,
  2273.019287109375,
  2241.18212890625,
  2203.78173828125,
  2164.943359375,
  2135.482421875,
  2098.81396484375,
  2063.52783203125,
  2029.579345703125,
  1992.4945068359375,
  1955.4969482421875,
  1922.6810302734375,
  1889.75048828125,
  1857.3846435546875,
  1818.62353515625,
  1785.3614501953125,
  1750.5054931640625,
  1718.0311279296875,
  1683.3194580078125,
  1649.578369140625,
  1614.3935546875,
  1575.2589111328125,
  1544.287841796875,
  1510.8096923828125,
  1475.779052734375,
  1443.775390625,
  1414.1778564453125,
  1380.3551025390625,
  1344.8077392578125,
  1312.089111328125,
  1279.7847900390625,
  1248.6295166015625,
  1217.0859375,
  1180.9456787109375,
  1155.2806396484375,
  1125.015380859375,
  1099.8720703125,
  1070.0064697265625,
  1041.020751953125,
  1014.7119750976562,
  986.2686767578125,
  959.3794555664062,
  935.0941772460938,
  909.951904296875,
  883.1526489257812,
  855.6653442382812,
  828.5133056640625,
  801.3621215820312,
  774.6409912109375,
  747.5872192382812,
  722.4879760742188,
  697.6157836914062,
  677.3233032226562,
  652.6466674804688,
  626.4276733398438,
  601.671630859375,
  580.1514282226562,
  558.9360961914062,
  539.7982788085938,
  518.058837890625,
  494.1725769042969,
  473.908203125,
  450.6574401855469,
  427.4228210449219,
  407.8333435058594,
  385.520263671875,
  369.1092834472656,
  346.7403259277344,
  334.2724609375,
  315.586181640625,
  299.3265380859375,
  282.21661376953125,
  265.4549255371094,
  250.22999572753906,
  236.50067138671875,
  221.37860107421875,
  207.23692321777344,
  192.8123016357422,
  181.2073974609375,
  169.5457305908203,
  158.15614318847656,
  146.18380737304688,
  134.43270874023438,
  124.99821472167969,
  116.7092514038086,
  107.42877197265625,
  98.04393768310547,
  89.8097915649414,
  80.06987762451172,
  74.47185516357422,
  68.15503692626953,
  61.96371078491211,
  56.354034423828125,
  52.5781135559082,
  48.67445373535156,
  45.53632354736328,
  42.20598602294922,
  39.4415283203125,
  36.46059799194336,
  34.38557434082031,
  32.91263198852539,
  31.126068115234375,
  29.82269287109375,
  29.08582878112793,
  27.707895278930664,
  26.97379493713379,
  26.495267868041992,
  26.45466423034668,
  25.580915451049805,
  25.453462600708008,
  25.348770141601562,
  24.757976531982422,
  25.330747604370117,
  24.738906860351562,
  24.67300796508789,
  24.469894409179688,
  24.374359130859375,
  24.60072898864746,
  24.245256423950195,
  24.414562225341797,
  24.620513916015625,
  24.377910614013672,
  24.04123878479004,
  24.57693099975586,
  24.388919830322266,
  24.219173431396484,
  24.378971099853516,
  24.30501365661621,
  24.48757553100586,
  24.383054733276367,
  24.285554885864258,
  24.646129608154297,
  24.462970733642578,
  24.12499237060547,
  24.14937973022461,
  24.333009719848633,
  24.554513931274414,
  24.275711059570312,
  24.358694076538086,
  24.228403091430664,
  24.289846420288086,
  24.76703453063965,
  24.149415969848633,
  24.046483993530273,
  24.071815490722656,
  23.92632293701172,
  24.532142639160156,
  24.749914169311523,
  24.25486946105957,
  24.304670333862305,
  24.236284255981445,
  23.766464233398438]}

loss_df = pd.DataFrame(model.history.history)

loss_df.plot()

<matplotlib.axes._subplots.AxesSubplot at 0x7fa25bddc850>

model.evaluate(X_test, y_test, verbose = 0)

25.009714126586914

model.evaluate(X_train,y_train, verbose = 3 )

24.01677703857422

test_predictions = model.predict(X_test)

test_predictions

array([[405.06985],
       [623.33136],
       [591.92316],
       [572.01245],
       [366.38934],
       [578.968  ],
       [514.83344],
       [458.8928 ],
       [549.03033],
       [447.30484],
       [611.57715],
       [548.73895],
       [418.88043],
       [408.64896],
       [651.0408 ],
       [437.08276],
       [508.2425 ],
       [659.7209 ],
       [662.3756 ],
       [565.32855],
       [334.00137],
       [444.61594],
       [382.30347],
       [378.53406],
       [566.3838 ],
       [610.38074],
       [532.1772 ],
       [427.64233],
       [655.23486],
       [413.8644 ],
       [442.42065],
       [484.9126 ],
       [438.1969 ],
       [681.6411 ],
       [424.5428 ],
       [417.465  ],
       [501.71875],
       [550.2641 ],
       [509.51538],
       [395.13196],
       [618.3842 ],
       [416.41095],
       [604.24146],
       [445.74643],
       [501.7747 ],
       [581.5117 ],
       [668.8    ],
       [490.18997],
       [318.35724],
       [485.21155],
       [517.05194],
       [381.57352],
       [541.70087],
       [408.3758 ],
       [641.3148 ],
       [490.86148],
       [627.7333 ],
       [626.79803],
       [446.85458],
       [484.48703],
       [490.9208 ],
       [474.32666],
       [682.5859 ],
       [403.0684 ],
       [700.93396],
       [586.18933],
       [582.9135 ],
       [537.6466 ],
       [484.3618 ],
       [516.34717],
       [361.00427],
       [540.5449 ],
       [570.4192 ],
       [528.2499 ],
       [453.5069 ],
       [530.9317 ],
       [506.99905],
       [443.11868],
       [543.15704],
       [640.5515 ],
       [465.79477],
       [567.0515 ],
       [690.5667 ],
       [458.52057],
       [708.70984],
       [472.4834 ],
       [403.06332],
       [584.896  ],
       [436.54315],
       [488.75186],
       [616.83264],
       [439.26276],
       [455.11356],
       [435.04376],
       [506.70016],
       [608.2418 ],
       [321.40756],
       [436.0566 ],
       [536.1338 ],
       [518.4238 ],
       [604.8458 ],
       [525.23724],
       [333.91547],
       [575.959  ],
       [431.56796],
       [562.3486 ],
       [513.1956 ],
       [390.8673 ],
       [566.04395],
       [454.5859 ],
       [448.2718 ],
       [640.7583 ],
       [524.1675 ],
       [550.4805 ],
       [417.4784 ],
       [478.5295 ],
       [586.31744],
       [667.1741 ],
       [700.2619 ],
       [659.19116],
       [560.35895],
       [502.96515],
       [390.13855],
       [281.11826],
       [479.40378],
       [616.1689 ],
       [373.0969 ],
       [512.0524 ],
       [510.81815],
       [493.2759 ],
       [480.24887],
       [423.57587],
       [493.24786],
       [471.26056],
       [600.2448 ],
       [573.37756],
       [414.75076],
       [630.44574],
       [466.16125],
       [564.1327 ],
       [405.47202],
       [531.9684 ],
       [572.4318 ],
       [356.86008],
       [549.6302 ],
       [603.15405],
       [384.08636],
       [542.1864 ],
       [562.39026],
       [452.67438],
       [632.0046 ],
       [372.03677],
       [474.04315],
       [528.662  ],
       [372.08386],
       [461.02655],
       [436.09967],
       [498.28693],
       [345.89633],
       [395.07632],
       [604.38116],
       [506.4469 ],
       [468.4045 ],
       [490.15884],
       [535.0751 ],
       [344.2807 ],
       [512.1184 ],
       [250.57906],
       [503.88028],
       [540.8969 ],
       [489.16257],
       [470.87903],
       [392.33386],
       [416.04172],
       [549.37286],
       [475.5845 ],
       [579.7455 ],
       [489.563  ],
       [600.7843 ],
       [546.7931 ],
       [541.759  ],
       [500.3749 ],
       [645.6394 ],
       [560.214  ],
       [577.737  ],
       [443.9203 ],
       [415.30664],
       [419.8243 ],
       [568.92145],
       [608.8123 ],
       [437.54663],
       [487.76324],
       [587.55054],
       [525.1769 ],
       [357.14728],
       [645.2205 ],
       [527.82904],
       [336.99448],
       [492.4843 ],
       [409.9426 ],
       [606.402  ],
       [346.31693],
       [522.2128 ],
       [404.4764 ],
       [258.1915 ],
       [519.69073],
       [340.679  ],
       [361.831  ],
       [576.2177 ],
       [416.46265],
       [550.78394],
       [520.87897],
       [510.55246],
       [324.53278],
       [403.93765],
       [602.2005 ],
       [617.6617 ],
       [603.0705 ],
       [566.08905],
       [473.21838],
       [460.15936],
       [508.22592],
       [445.69302],
       [511.10815],
       [502.17154],
       [399.92477],
       [605.32947],
       [257.9128 ],
       [628.05505],
       [588.93463],
       [326.92877],
       [479.34045],
       [595.14575],
       [378.18112],
       [459.9446 ],
       [324.79074],
       [518.4878 ],
       [409.40533],
       [556.04675],
       [641.694  ],
       [536.6195 ],
       [502.97104],
       [634.8414 ],
       [515.03577],
       [531.5037 ],
       [518.6183 ],
       [457.23022],
       [505.62198],
       [460.8963 ],
       [591.5011 ],
       [465.702  ],
       [426.96902],
       [541.3878 ],
       [493.59402],
       [679.4696 ],
       [372.551  ],
       [551.4076 ],
       [577.7646 ],
       [433.75317],
       [542.82   ],
       [585.79913],
       [579.0469 ],
       [720.7482 ],
       [432.5301 ],
       [398.3806 ],
       [313.78818],
       [448.07065],
       [387.87238],
       [542.9997 ],
       [522.5298 ],
       [564.28656],
       [447.80322],
       [534.2046 ],
       [381.5742 ],
       [501.3084 ],
       [636.9072 ],
       [496.34296],
       [568.1877 ],
       [469.90048],
       [273.087  ],
       [517.1568 ],
       [621.2014 ],
       [350.29446],
       [450.2327 ],
       [499.081  ],
       [542.76935],
       [611.6272 ],
       [387.9596 ],
       [449.1869 ],
       [482.20422],
       [598.1141 ],
       [499.1719 ],
       [321.23654],
       [554.9411 ],
       [444.4191 ],
       [528.82935],
       [515.10876],
       [609.4433 ],
       [416.73828],
       [410.65424]], dtype=float32)

test_predictions = pd.Series(test_predictions.reshape(300,))

pred_df = pd.DataFrame(y_test, columns=['Test True Y'])

pred_df = pd.concat([pred_df, test_predictions], axis = 1)

pred_df

pred_df.columns = ['Test True Y', 'Model Predictions']

plt.figure(figsize=(10, 6))
sns.scatterplot(data = pred_df, x='Test True Y', y='Model Predictions')

<matplotlib.axes._subplots.AxesSubplot at 0x7fa255ee1310>

from sklearn.metrics import mean_absolute_error, mean_squared_error

mean_absolute_error(pred_df['Test True Y'], pred_df['Model Predictions'])

3.9980359288852494

df.describe()

mean_squared_error(pred_df['Test True Y'], pred_df['Model Predictions'])

24.94500463893638

# to get the root mean squared error, then raise to the power by 0.5
mean_squared_error(pred_df['Test True Y'], pred_df['Model Predictions']) ** 0.5

4.994497436072661

# predicting on brand new data set
new_gem =[[ 998, 1000]]

scaler.transform(new_gem)

array([[0.14117652, 0.53968792]])

new_gem = scaler.transform(new_gem)

model.predict(new_gem)

array([[419.4626]], dtype=float32)

# to save a model that is doing well, then from tensorflow models import load model

from tensorflow.keras.models import load_model

model.save('new_mygem')

INFO:tensorflow:Assets written to: new_mygem/assets

# ** LAST CELL ** #

	price	feature1	feature2
0	461.527929	999.787558	999.766096
1	548.130011	998.861615	1001.042403
2	410.297162	1000.070267	998.844015
3	540.382220	999.952251	1000.440940
4	546.024553	1000.446011	1000.338531

	Test True Y	0
0	402.296319	405.069855
1	624.156198	623.331360
2	582.455066	591.923157
3	578.588606	572.012451
4	371.224104	366.389343
...	...	...
295	525.704657	528.829346
296	502.909473	515.108765
297	612.727910	609.443298
298	417.569725	416.738281
299	410.538250	410.654236

	price	feature1	feature2
count	1000.000000	1000.000000	1000.000000
mean	498.673029	1000.014171	999.979847
std	93.785431	0.974018	0.948330
min	223.346793	997.058347	996.995651
25%	433.025732	999.332068	999.316106
50%	502.382117	1000.009915	1000.002243
75%	564.921588	1000.637580	1000.645380
max	774.407854	1003.207934	1002.666308

Consulting - Quality

TENSOR FLOW

WITH FAKE DATASET