forecast
Predict future values from time series data using point or probabilistic outputs. Supports univariate and multivariate series with Chronos2 or TiRex models.
Instructions
Perform time series forecasting using FAIM platform. Supports both point forecasting (single value) and probabilistic forecasting (confidence intervals). Can handle univariate and multivariate time series data. Currently supported models: Chronos2 (default, recommended for multivariate) and TiRex (fast, univariate only).
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| model | Yes | The forecasting model to use. Chronos2: State-of-the-art, supports univariate/multivariate, custom quantiles. TiRex: Fast alternative for univariate only, uses fixed quantiles [0.1,0.2,...,0.9], custom quantiles parameter ignored. | |
| x | No | Time series data to forecast from. MUST be an array, NOT a string. Can be a 1D array [1,2,3,4,5], 2D array [[1,2],[3,4]] (multiple series/batch or multivariate per model), or 3D array [[[1],[2]]] (batch, sequence, features). Never pass x as a JSON string - always pass as an actual array. | |
| horizon | Yes | Number of time steps to forecast into the future. Must be a positive integer. Example: 10 means predict the next 10 steps. | |
| output_type | No | Type of forecast output. "point" = single value per step (fastest). "quantiles" = confidence intervals (use for uncertainty estimation). Default: "point". | |
| quantiles | No | Custom quantile levels to compute (only used with output_type="quantiles" and Chronos2 model). For TiRex, this parameter is ignored and fixed quantiles [0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9] are always returned. Values must be between 0 and 1. Example: [0.1, 0.5, 0.9] for 10th, 50th, 90th percentiles. | |
| is_multivariate | No | For 2D input arrays only with Chronos2: interpret as multivariate time series (true) or batch of univariate series (false, default). Ignored for 1D arrays, 3D arrays, and TiRex model. |