-
Notifications
You must be signed in to change notification settings - Fork 0
/
thirty_seconds_prediction.py
139 lines (105 loc) · 4.72 KB
/
thirty_seconds_prediction.py
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
import pandas as pd # pandas is a dataframe library
import numpy as np
import matplotlib.pyplot as plt # matplotlib.pyplot plots data
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense, Dropout, LSTM, CuDNNLSTM
from keras.callbacks import EarlyStopping
from keras.utils import plot_model
from keras import backend as K
import os
os.environ["PATH"] += os.pathsep + 'C:/Program Files (x86)/Graphviz2.38/bin/'
OUTPUTS = ['inlet_probe_1_rack_1', 'inlet_probe_2_rack_1', 'inlet_probe_3_rack_1', 'inlet_probe_4_rack_1',
'inlet_probe_1_rack_2', 'inlet_probe_2_rack_2', 'inlet_probe_3_rack_2', 'inlet_probe_4_rack_2',
'outlet_probe_1_rack_1', 'outlet_probe_1_rack_2', 'outlet_probe_1_room',
'outlet_probe_2_room', 'outlet_probe_3_room', 'outlet_probe_4_room']
def r2_keras(y_true, y_pred):
SS_res = K.sum(K.square(y_true - y_pred))
SS_tot = K.sum(K.square(y_true - K.mean(y_true)))
return (1 - SS_res/(SS_tot + K.epsilon()))
def plot_predictions(real, predicted, column):
fig, ax = plt.subplots()
line1, = ax.plot(np.arange(30), real)
line2, = ax.plot(np.arange(30), predicted)
line1.set_label('Real temperature')
line2.set_label('Predicted temperature')
ax.set_title(column)
ax.set_xlabel('Second')
ax.set_ylabel('Temperature')
ax.legend()
plt.savefig('./images/thirty_seconds_data/' + column + '.png')
plt.close(fig)
def plot_learning_rates(hist, column, score):
epoch_list = list(range(1, len(hist.history[score]) + 1)) # values for x axis [1, 2, .., # of epochs]
fig, ax = plt.subplots()
line1, = ax.plot(epoch_list, hist.history[score])
line2, = ax.plot(epoch_list, hist.history['val_' + score])
line1.set_label('Training ' + score)
line2.set_label('Validation ' + score)
ax.set_title(column)
ax.set_xlabel('Epoch')
ax.set_ylabel('Score')
ax.legend()
plt.savefig('./images/thirty_seconds_data/' + score + '_' + column + '.png')
plt.close(fig)
prediction_index = 0
for column in OUTPUTS:
df = pd.read_csv("E:/data/thirty_seconds_data/" + column + ".csv")
print(df.shape)
X = df[df.columns[0:65]].values # predictor feature columns (10 X m)
y = df[df.columns[65:95]].values # predicted class column (1 X m)
split_test_size = 0.2
# normalize data
scaler = MinMaxScaler()
x_scaled = scaler.fit_transform(X)
y_scaled = scaler.fit_transform(y)
X_train, X_test, y_train, y_test = train_test_split(x_scaled, y_scaled, test_size=split_test_size, random_state=42)
X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1))
X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))
model = Sequential()
# input shape = (6,)
model.add(LSTM(65, input_shape=(X_train.shape[1], 1), return_sequences=True))
add(Dropout(0.2))
model.add(LSTM(130, return_sequences=True))
model.add(Dropout(0.2))
model.add(LSTM(130))
model.add(Dropout(0.2))
model.add(Dense(30, activation='linear'))
# compile
model.compile(loss='mean_absolute_error',
optimizer='adam',
metrics=[r2_keras])
# save model to png
plot_model(model, to_file="./images/thirty_seconds_data.png", show_shapes=True, show_layer_names=True)
# train
BATCH_SIZE = 750
EPOCHS = 5
cbk_early_stopping = EarlyStopping(monitor='val_r2_keras', mode='max')
hist = model.fit(X_train, y_train, BATCH_SIZE, epochs=EPOCHS,
validation_data=(X_test, y_test),
callbacks=[cbk_early_stopping])
# evaluate the model with the test data to get the scores on "real" data
score = model.evaluate(X_test, y_test, verbose=2)
print("All score", score)
print('Test loss:', score[0])
print('Test r2: ', score[1])
# do predictions
print("Prediction")
predicted_temperature = model.predict(X_test)
predicted_temperature = scaler.inverse_transform(predicted_temperature)
real_temperature = scaler.inverse_transform(y_test)
# plot prediction
print('Predicted: ', predicted_temperature[0])
print('Real: ', real_temperature[0])
file = open("./results/thirty_seconds_data/result.txt", "a")
file.write("Column: " + column + "\n")
file.write("Test loss, Test r2" + "\n")
file.write("Test loss, Test r2" + "\n")
file.write(str(score[0]) + " " + str(score[1]) + "\n")
file.write("\n")
file.close()
plot_predictions(real_temperature[prediction_index], predicted_temperature[prediction_index], column)
prediction_index = prediction_index + 1
plot_learning_rates(hist, column, 'r2_keras')
plot_learning_rates(hist, column, 'loss')