def test_base_parameters_split_all_0(self):
X = None
y = numpy.arange(5) * 100
weights = numpy.arange(5) * 1000
bs = BaseTimeSeries(past=2, use_all_past=True)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqualArray(y[0:-2], nx[:, 0])
self.assertEqualArray(y[1:-1], nx[:, 1])
self.assertEqualArray(y[2:].reshape((3, 1)), ny)
self.assertEqualArray(weights[1:-1], nw)
def test_base_parameters_split_all_0_same(self):
X = None
y = numpy.arange(5).astype(numpy.float64) * 100
weights = numpy.arange(5).astype(numpy.float64) * 1000
bs = BaseTimeSeries(past=2, use_all_past=True)
nx, ny, nw = build_ts_X_y(bs, X, y, weights, same_rows=True)
self.assertEqualArray(y[0:-2], nx[2:, 0])
self.assertEqualArray(y[1:-1], nx[1:-1, 1])
self.assertEqualArray(y[2:].reshape((3, 1)), ny[2:])
self.assertEqualArray(weights, nw)
def test_base_parameters_split0_1(self):
X = None
y = numpy.arange(5) * 100
weights = numpy.arange(5) + 1000
bs = BaseTimeSeries(past=1)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqual(nx.shape, (4, 1))
self.assertEqualArray(y[0:-1], nx[:, 0])
self.assertEqualArray(y[1:].reshape((4, 1)), ny)
self.assertEqualArray(weights[:-1], nw)
def test_base_parameters_split2(self):
X = numpy.arange(10).reshape(5, 2)
y = numpy.arange(5) * 100
weights = numpy.arange(5) * 1000
bs = BaseTimeSeries(past=2, delay2=3)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqualArray(X[1:-2], nx[:, :2])
self.assertEqualArray(y[0:-3], nx[:, 2])
self.assertEqualArray(y[1:-2], nx[:, 3])
self.assertEqualArray(numpy.array([[200, 300], [300, 400]]), ny)
self.assertEqualArray(weights[1:-2], nw)
def test_base_parameters_split1(self):
X = numpy.arange(10).reshape(5, 2)
y = numpy.arange(5) * 100
weights = numpy.arange(5) * 1000
bs = BaseTimeSeries(past=2)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqualArray(X[1:-1], nx[:, :2])
self.assertEqualArray(y[0:-2], nx[:, 2])
self.assertEqualArray(y[1:-1], nx[:, 3])
self.assertEqualArray(y[2:].reshape((3, 1)), ny)
self.assertEqualArray(weights[1:-1], nw)
def test_base_parameters_split_all_1(self):
X = numpy.arange(10).reshape(5, 2)
y = numpy.arange(5) * 100
weights = numpy.arange(5) * 1000
bs = ARTimeSeriesRegressor(past=2, use_all_past=True)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqualArray(X[0:-2], nx[:, :2])
self.assertEqualArray(X[1:-1], nx[:, 2:4])
self.assertEqualArray(y[0:-2], nx[:, 4])
self.assertEqualArray(y[1:-1], nx[:, 5])
self.assertEqualArray(y[2:].reshape((3, 1)), ny)
self.assertEqualArray(weights[1:-1], nw)
def test_base_parameters_split_all_2(self):
X = numpy.arange(10).reshape(5, 2)
y = numpy.arange(5) * 100
weights = numpy.arange(5) * 1000
bs = ARTimeSeriesRegressor(past=2, delay2=3, use_all_past=True)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqualArray(X[0:-3], nx[:, :2])
self.assertEqualArray(X[1:-2], nx[:, 2:4])
self.assertEqualArray(y[0:-3], nx[:, 4])
self.assertEqualArray(y[1:-2], nx[:, 5])
self.assertEqualArray(numpy.array([[200, 300], [300, 400]]), ny)
self.assertEqualArray(weights[1:-2], nw)
def test_base_parameters_split0(self):
X = numpy.arange(20).reshape((10, 2))
y = numpy.arange(10) * 100
bs = BaseTimeSeries(past=2)
nx, ny, _ = build_ts_X_y(bs, X, y)
for d in range(0, 5):
proc = TimeSeriesDifference(d)
proc.fit(nx, ny)
px, py = proc.transform(nx, ny)
self.assertEqualArray(px[-1, :], nx[-1, :])
rev = proc.get_fct_inv()
ppx, ppy = rev.transform(px, py)
self.assertEqualArray(nx, ppx)
self.assertEqualArray(ny, ppy)
def test_fit_predict(self):
X = None
y = numpy.arange(5) * 100
weights = numpy.arange(5) * 1000
bs = ARTimeSeriesRegressor(past=2)
nx, ny, nw = build_ts_X_y(bs, X, y, weights)
self.assertEqualArray(y[0:-2], nx[:, 0])
self.assertEqualArray(y[1:-1], nx[:, 1])
self.assertEqualArray(y[2:].reshape((3, 1)), ny)
self.assertEqualArray(weights[1:-1], nw)
nx = nx.astype(numpy.float64)
ny = ny.astype(numpy.float64)
bs.fit(nx, ny)
pred = bs.predict(nx, ny)
self.assertEqual(pred.shape, ny.shape)
