TimeSequencePredictor


TimeSequencePredictor

TimeSequencePredictor can be used to train a pipeline (including feature engineering and model) for automated time series forecasting in a distributed way. AutoML is applied for searching the best set of features as well as model hyper-parameters.

Methods

__init__

tsp = TimeSequencePredictor(name="automl",
                            logs_dir="~/zoo_automl_logs",
                            future_seq_len=1,
                            dt_col="datetime",
                            target_col="value",
                            extra_features_col=None,
                            drop_missing=True,)

Arguments

fit

Train a pipeline for time series forecasting. It will return a TimeSequencePipeline object.

tsp.fit(self,
        input_df,
        validation_df=None,
        metric="mse",
        recipe=SmokeRecipe(),
        mc=False,
        resources_per_trial={"cpu": 2},
        distributed=False
        )

Arguments