Computerized methods and systems for machine-learned multi-output multi-step forecasting of time-series data

A non-transitory computer-readable medium comprising instructions including:

• displaying, on a user device, a graphical user interface of a web portal;
• and
• in response to receiving a forecasting request from the user device via the graphical user interface of the web portal:
• determining a measure of uncertainty associated with an input vector specified by the forecasting request, wherein the measure of uncertainty includes a compact representation of uncertainty associated with future time intervals specified by the forecasting request;
• obtaining a set of historical data from a time-series data store;
• generating a forecast model using the set of historical data, the input vector, and the measure of uncertainty to predict outcomes incrementally for the future time intervals, wherein:
• the forecast model is configured to generate a mean vector and a predicted uncertainty measure,
• the predicted uncertainty measure is generated based on the compact representation of uncertainty,
• the mean vector includes mean values corresponding respectively to the future time intervals, and
• the compact representation of uncertainty is carried forward for the future time intervals;
• determining a predicted outcome using the forecast model at an end of the future time intervals;
• and
• in response to the predicted outcome exceeding a threshold, updating, via the web portal, the graphical user interface at the user device to display:
• the mean vector and a magnitude of the predicted uncertainty measure, and
• a communication indicating a proposed action associated with the input vector, the set of historical data, the mean vector, and the magnitude of the predicted uncertainty measure.

A computer-implemented method comprising:

• displaying, on a user device, a graphical user interface of a web portal;
• and
• in response to receiving a forecasting request from the user device via the graphical user interface of the web portal:
• determining a measure of uncertainty associated with an input vector specified by the forecasting request, wherein the measure of uncertainty includes a compact representation of uncertainty associated with future time intervals specified by the forecasting request;
• obtaining a set of historical data from a time-series data store;
• generating a forecast model using the set of historical data, the input vector, and the measure of uncertainty to predict outcomes incrementally for the future time intervals, wherein:
• the forecast model is configured to generate a mean vector and a predicted uncertainty measure,
• the predicted uncertainty measure is generated based on the compact representation of uncertainty,
• the mean vector includes mean values corresponding respectively to the future time intervals, and
• the compact representation of uncertainty is carried forward for the future time intervals;
• determining a predicted outcome using the forecast model at an end of the future time intervals;
• and
• in response to the predicted outcome exceeding a threshold, updating, via the web portal, the graphical user interface at the user device to display:
• the mean vector and a magnitude of the predicted uncertainty measure, and
• a communication indicating a proposed action associated with the input vector, the set of historical data, the mean vector, and the magnitude of the predicted uncertainty measure.