LSTM-FCN PyTorch

PyTorch implementation of univariate time series classification model introduced in Karim, F., Majumdar, S., Darabi, H. and Chen, S., 2017. LSTM fully convolutional networks for time series classification. IEEE access, 6, pp.1662-1669.

Dependencies

numpy==1.23.5
torch==1.13.1
scikit-learn==1.1.3
plotly==5.11.0
kaleido==0.2.1

Usage

import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

from lstm_fcn_pytorch.model import Model
from lstm_fcn_pytorch.plots import plot

# Generate the data
N = 600    # number of time series
L = 100    # length of each time series
x = np.zeros((N, L))
t = np.linspace(0, 1, L)
c = np.cos(2 * np.pi * (10 * t - 0.5))
s = np.sin(2 * np.pi * (20 * t - 0.5))
x[:N // 3] = 10 + 10 * c + 5 * np.random.normal(size=(N // 3, L))
x[N // 3: 2 * N // 3] = 10 + 10 * s + 5 * np.random.normal(size=(N // 3, L))
x[2 * N // 3:] = 10 + 10 * c + 10 * s + 5 * np.random.normal(size=(N // 3, L))
y = np.concatenate([0 * np.ones(N // 3), 1 * np.ones(N // 3), 2 * np.ones(N // 3)])

# Split the data
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.3, stratify=y)

# Fit the model
model = Model(
    x=x_train,
    y=y_train,
    units=[5, 5],
    filters=[4, 4],
    kernel_sizes=[3, 3],
    dropout=0.2,
)

model.fit(
    learning_rate=0.001,
    batch_size=32,
    epochs=100,
    verbose=True
)

# Evaluate the model
yhat_train = model.predict(x_train)
yhat_test = model.predict(x_test)
print('Training accuracy: {:.6f}'.format(accuracy_score(y_train, yhat_train)))
print('Test accuracy: {:.6f}'.format(accuracy_score(y_test, yhat_test)))

# Plot the results
fig = plot(x=x_test, y=yhat_test)
fig.write_image('results.png', scale=4, height=900, width=700)