skits Multi Time Series

skitsmultits.py

from sklearn.linear_model import LinearRegression
from sklearn.pipeline import FeatureUnion

from skits.feature_extraction import AutoregressiveTransformer
from skits.pipeline import ForecasterPipeline
from skits.preprocessing import ReversibleImputer

df = pd.DataFrame(
    {
        "date": [1, 1, 1, 2, 2, 2, 3, 3 ,3, 4, 4, 4, 5, 5, 5],
        "ts_name": ["A", "B", "C", "A", "B", "C","A", "B", "C","A", "B", "C","A", "B", "C"],
        "Y_t": [1.1, 2.3, 3.1, 1.2, 2.2, 3.3, 1.05, 2.25, 3.35, np.nan, np.nan, np.nan, np.nan, np.nan, np.nan,],
    }
)

forecast_date_start = 4

pipeline = ForecasterPipeline(
    [
        (
            "features",
            FeatureUnion(
                [("ar_transformer", AutoregressiveTransformer(num_lags=2))]
            ),
        ),
        ("post_lag_imputer", ReversibleImputer()),
    ]
)

Xts = []
yts = []

for ts_name, group in df.query('date < @forecast_date_start').groupby("ts_name"):

    y = group["Y_t"].values
    X = y[:, np.newaxis].copy()

    Xt = pipeline.fit_transform(X, y)
    yt = pipeline.transform_y(y)

    if yt.ndim == 1:
        yt = yt[:, np.newaxis]

    Xts.append(Xt)
    yts.append(yt)

Xt = np.vstack(Xts)
yt = np.vstack(yts)

estimator = LinearRegression()
# Forecast

start_idx = 3
end_idx = 5

for ts_name, group in df.groupby("ts_name"):
    y = group["Y_t"].values
    for idx in range(start_idx, end_idx):
        X = y[:(idx + 1), np.newaxis].copy()
        Xt = pipeline.transform(X)
        y[idx] = estimator.predict(Xt[-1:, ])
        df.loc[(df.date == idx+1) & (df.ts_name == ts_name), 'Y_t'] = y[idx]

df.groupby(['date','ts_name']).last()['Y_t'].unstack().plot()
plt.axvline(3);