[ANN] SimpleRNN (2)

데이터 전처리

 

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.layers import SimpleRNN, LSTM, Dense
from tensorflow.keras import Sequential

# data 생성
x = np.arange(0, 100, 0.1)
y = 0.5*np.sin(2*x) - np.cos(x/2.0)

seq_data = y.reshape(-1,1)

print(seq_data.shape)
print(seq_data[:5])

(1000, 1)
[[-1.        ]
 [-0.89941559]
 [-0.80029499]
 [-0.70644984]
 [-0.62138853]]

plt.grid()
plt.title('0.5*sin(2x)-cos(x/2)')
plt.xlabel('time')
plt.ylabel('amplitude')
plt.plot(seq_data)

plt.show()

def seq2dataset(seq, window, horizon):
    X = []
    Y = []

    for i in range(len(seq)-(window+horizon)+1):
        x = seq[i:(i+window)]
        y = (seq[i+window+horizon-1])

        X.append(x)
        Y.append(y)

    return np.array(X), np.array(Y)
    
w = 20  # window size
h = 1  # horizon factor

X, Y = seq2dataset(seq_data, w, h)

print(X.shape, Y.shape)

(980, 20, 1) (980, 1)

split_ratio = 0.8

split = int(split_ratio*len(X))

x_train = X[0:split]
y_train = Y[0:split]

x_test = X[split:]
y_test = Y[split:]

print(x_train.shape, y_train.shape, x_test.shape, y_test.shape)

(784, 20, 1) (784, 1) (196, 20, 1) (196, 1)

 

모델 생성

 

model = Sequential()

model.add(SimpleRNN(units=128, activation='tanh', input_shape=x_train[0].shape))

model.add(Dense(1))

model.summary()

Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 simple_rnn (SimpleRNN)      (None, 128)               16640     
                                                                 
 dense (Dense)               (None, 1)                 129       
                                                                 
=================================================================
Total params: 16,769
Trainable params: 16,769
Non-trainable params: 0
_________________________________________________________________

from datetime import datetime

model.compile(loss='mse', optimizer='adam', metrics=['mae'])

start_time = datetime.now()

hist = model.fit(x_train, y_train, 
                 epochs=100, 
                 validation_data=(x_test, y_test))

end_time = datetime.now()

print('Elapsed Time => ', end_time-start_time)

plt.title('Loss Trend')
plt.plot(hist.history['loss'], label='loss')
plt.plot(hist.history['val_loss'], label='val_loss')
plt.xlabel('epoch')
plt.ylabel('loss')
plt.grid()
plt.legend(loc='best')

plt.show()

pred = model.predict(x_test)

print(pred.shape)

7/7 [==============================] - 0s 4ms/step
(196, 1)

rand_idx = np.random.randint(0, len(y_test), size=5)

print('random idx = ',rand_idx, '\n')

print('pred  = ', pred.flatten()[rand_idx])
print('label = ', y_test.flatten()[rand_idx])

rand_idx = np.random.randint(0, len(y_test), size=5)

print('\n\nrandom idx = ',rand_idx, '\n')

print('pred  = ', pred.flatten()[rand_idx])
print('label = ', y_test.flatten()[rand_idx])

random idx =  [  9 152 148  25  84] 

pred  =  [ 0.6421721  1.0176921  1.3699332  1.1660998 -0.4494172]
label =  [ 0.6363903   0.99169316  1.36121199  1.14384382 -0.41652189]

random idx =  [152 181 100  75  14] 

pred  =  [ 1.0176921  -0.13476941 -0.8855729  -1.1017574   1.1269114 ]
label =  [ 0.99169316 -0.12930711 -0.83954079 -1.06389327  1.11572477]

plt.plot(pred, label='prediction')
plt.plot(y_test, label='label')
plt.grid()
plt.legend(loc='best')

plt.show()