Time-Series models for multivariate and multistep forecasting, regression, and classification

A prototype framework for developing multi-step machine learnings.

Predictor that uses a configurable plugin-based predictive supervised learning model, to make forecasts for a configurable time horizon in a timeseries, using heterogeneus multivariate timeseries data as input, the input data needs to be aligned with the timeseries to be used as training signals. Includes 4 built-in deep-learning predictive models.