The series forecasting functions help you to develop forecasting models using statistical models, seasonal data and historical data.
Functions:
- TD_ARIMAFORECAST
- Uses ARIMA algorithm to forecast user-defined number of periods into the future beyond the last observed sample point in the model. This function only outputs a primary result set, containing forecasted value.
- TD_DTW
- Finds the optimal, or a close to optimal warp path between two series depending on the search radius.
- TD_HOLT_WINTERS_FORECASTER
- Uses the Holt Winters model to forecast user-defined number of periods into the future beyond the last observed sample point in the model. This function outputs up to four result sets.
- TD_KALMAN_FORECASTER
- Generates a forecast and related metrics in secondary output layers. This function extends the linear regression model by accepting time-varying regression coefficients.
- TD_MAMEAN
- Uses available historical data to forecast activity user-defined number of periods into the future.
- TD_SIMPLEEXP
- Uses a forecasting model that uses the level modeling component to accomplish the forecasting of the original series.
- TD_SMOOTHMA_FORECASTER
- Applies a user-selected smoothing function to an input series, creating new time series highlighting the trend along with a naïve forecast.