def test_temporal_k_fold():
# test length 1 series
splitter = TemporalKFold()
X = [rand(100, 10)]
y = [5]
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
X = [rand(100, 10)]
y = [rand(100)]
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
Xt = [rand(100, 10)]
Xc = [5]
X = TS_Data(Xt, Xc)
y = [rand(100)]
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
# test with lots of series
splitter = TemporalKFold()
Ns = 5
X = np.array([rand(100, 10)] * Ns)
y = rand(Ns)
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
X = np.array([rand(100, 10)] * Ns)
y = np.array([rand(100)] * Ns)
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
Xt = np.array([rand(100, 10)] * Ns)
Xc = rand(Ns)
X = TS_Data(Xt, Xc)
y = np.array([rand(100)] * Ns)
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
Xt = np.array([rand(100, 10)] * Ns)
Xc = rand(Ns)
X = TS_Data(Xt, Xc)
y = rand(Ns)
Xs, ys, cv = splitter.split(X, y)
check_folds(Xs, ys, cv)
model.add(Dense(n_classes, activation="softmax"))
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
return model
# load the data
data = load_watch()
X = data['X']
y = data['y']
# temporal splitting of data
splitter = TemporalKFold(n_splits=3)
Xs, ys, cv = splitter.split(X, y)
# create a segment learning pipeline
width = 100
pipe = Pype([('seg', SegmentX(order='C')),
('crnn',
KerasClassifier(build_fn=crnn_model,
epochs=1,
batch_size=256,
verbose=0))])
# create a parameter dictionary using the sklearn API
#
# you can also set a parameter to be always equal to another parameter, by setting its value to
# parameter name to track (this is an extension from sklearn)
print("N series in train: ", len(X_train))
print("N series in test: ", len(X_test))
print("N segments in train: ", pipe.N_train)
print("N segments in test: ", pipe.N_test)
print("Score: ", score)
# generate some predictions
ytr, ytr_p = pipe.transform_predict(X_train, y_train) # training predictions
yte, yte_p = pipe.transform_predict(X_test, y_test) # test predictions
xtr = np.arange(len(ytr)) # segment number
xte = np.arange(len(yte)) + len(xtr)
# plot the amazing results
plt.plot(xtr, ytr, '.', label="training")
plt.plot(xte, yte, '.', label="actual")
plt.plot(xte, yte_p, label="predicted")
plt.xlabel("Segment Number")
plt.ylabel("Target")
plt.legend()
plt.show()
# now try a cross validation
X = [np.arange(4000) / 100.]
y = [np.sin(X[0])]
tkf = TemporalKFold()
X, y, cv = tkf.split(X, y)
cv_scores = cross_validate(pipe, X, y, cv=cv, return_train_score=True)
print("CV Scores: ", pd.DataFrame(cv_scores))