Homec4science
Diffusion projectML (master) Browse

R2096/lstm/lstm.py
296769ee52f9master

Last Change
Raw File

lstm.py
View Options

# LSTM for international airline passengers problem with regression framing
import numpy
import scipy.io
import matplotlib.pyplot as plt
from pandas import read_csv
import math
import sys
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import mean_squared_error
# FUNCTION: convert an array of values into a dataset matrix
def create_dataset(dataset, seq_length=1):
dataX, dataY = [], []
for i in range(len(dataset)-seq_length):
dataX.append(dataset[i:(i+seq_length), 0])
dataY.append(dataset[i + seq_length, 0])
return numpy.array(dataX), numpy.array(dataY)
# FUNCTION: lstm
def lstm(fileName, ratioTrain = 0.67, seq_length = 1, nbNodeHiddenLayer = 4, epochs_=10, batch_size_=1, verbose_=2):
numpy.random.seed(7)
# load the dataset
dataframe = read_csv(fileName, usecols=[1], engine='python', header=None)
dataset = dataframe.values
dataset = dataset.astype('float32')
# normalize the dataset
scaler = MinMaxScaler(feature_range=(0, 1))
dataset = scaler.fit_transform(dataset)
# split into train and test sets
train_size = int(len(dataset) * ratioTrain)
test_size = len(dataset) - train_size
train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]
# reshape into X=t and Y=t+1
trainX, trainY = create_dataset(train, seq_length)
testX, testY = create_dataset(test, seq_length)
# reshape input to be [samples, time steps, features]
trainX = numpy.reshape(trainX, (trainX.shape[0], 1, trainX.shape[1]))
testX = numpy.reshape(testX, (testX.shape[0], 1, testX.shape[1]))
# create and fit the LSTM network
model = Sequential()
model.add(LSTM(nbNodeHiddenLayer, input_shape=(1, seq_length)))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')
model.fit(trainX, trainY, epochs = epochs_, batch_size = batch_size_, verbose = verbose_)
# make predictions
Otrain = model.predict(trainX)
Otest = model.predict(testX)
# invert predictions
Otrain = scaler.inverse_transform(Otrain)
trainY = scaler.inverse_transform([trainY])
Otest = scaler.inverse_transform(Otest)
testY = scaler.inverse_transform([testY])
dataset2 = scaler.inverse_transform(dataset)
# calculate root mean squared error
# trainScore = math.sqrt(mean_squared_error(trainY[0], Otrain[:,0]))
# testScore = math.sqrt(mean_squared_error(testY[0], Otest[:,0]))
trainScore = mean_squared_error(trainY[0], Otrain[:,0])/2
testScore = mean_squared_error(testY[0], Otest[:,0])/2
score = [trainScore, testScore];
# IO for Matlab
scipy.io.savemat('../output/dataset.mat', mdict={'dataset': dataset2})
scipy.io.savemat('../output/Otrain.mat', mdict={'Otrain': Otrain})
scipy.io.savemat('../output/Oval.mat', mdict={'Oval': Otest})
scipy.io.savemat('../output/score.mat', mdict={'score': score})
if __name__ == "__main__":
a = str(sys.argv[1])
b = float(sys.argv[2])
c = int(sys.argv[3])
d = int(sys.argv[4])
e = int(sys.argv[5])
f = int(sys.argv[6])
lstm(a,b,c,d,e,f)
c4science · Help