tw.forecasting
The Forecasting
class provides a comprehensive suite of methods for forecasting time-series data. It supports operations like model building, prediction generation, root cause analysis (RCA), and automatic forecasting, making it an essential tool for advanced data-driven forecasting tasks.
Class Initialization
from tangent_works import TangentWorks
# Initialize the TangentWorks instance
tw = TangentWorks()
# Access the Forecasting class
tw.forecasting
Methods
build_model
tangent_works.forecasting.build_model(configuration=Dict[str, Any], dataset=pd.DataFrame)
Builds a forecasting model using the provided configuration and dataset.
Parameters
configuration
(Dict[str, Any]): The configuration parameters for model building.dataset
(pd.DataFrame): The time-series dataset used to train the model.
Returns
ForecastingModel
: The constructed forecasting model.
Examples
model = tw.forecasting.build_model(
configuration = config,
dataset = data
)
predict
tangent_works.forecasting.predict(configuration=Dict[str, Any], dataset=pd.DataFrame, model=ForecastingModel)
Generates predictions using the provided configuration, dataset, and pre-trained forecasting model.
Parameters
configuration
(Dict[str, Any]): The configuration parameters for prediction.dataset
(pd.DataFrame): The input time-series data for generating predictions.model
(ForecastingModel): A previously built forecasting model.
Returns
pd.DataFrame
: The generated predictions.
Examples
predictions = tw.forecasting.predict(
configuration = config,
dataset = data,
model = model
)
rca
tangent_works.forecasting.rca(configuration=Dict[str, Any], dataset=pd.DataFrame, model=ForecastingModel)
Performs root cause analysis (RCA) on the dataset to identify potential causes of anomalies or variations.
Parameters
configuration
(Dict[str, Any]): Configuration parameters for performing RCA.dataset
(pd.DataFrame): The time-series dataset to analyze.model
(ForecastingModel): The forecasting model used for RCA.
Returns
Dict[int, pd.DataFrame]
: A dictionary mapping identified root causes to their corresponding model indexes.
Examples
rca_results = tw.forecasting.rca(
configuration = config,
dataset = data,
model = model
)
auto_forecast
tangent_works.forecasting.auto_forecast(configuration=Dict[str, Any], dataset=pd.DataFrame)
Automatically builds a forecasting model and generates predictions in a single step.
Parameters
configuration
(Dict[str, Any]): Configuration parameters for auto-forecasting.dataset
(pd.DataFrame): The input time-series dataset to analyze and forecast.
Returns
AutoForecastingResult
: Contains the built model and predictions.
Examples
auto_result = tw.forecasting.auto_forecast(
configuration = config,
dataset = data
)
Key Features
Modular Design: Separate methods for building models, generating predictions, and RCA.
Automation:
auto_forecast
combines model building and prediction into a single step.Flexibility: Accepts various configurations for custom forecasting workflows.
Dependencies
pandas: Used for handling time-series data.
Tangent Works core and business logic modules: Used for model building, validation, and forecasting.