| CPC G06F 40/40 (2020.01) | 21 Claims |
|
1. A computer-implemented method, comprising:
training a machine learning model to process time-series data received from a plurality of sources and generate natural language statistically supported alerts, wherein training the machine learning model, includes:
producing a tabular data set of time-series data;
generating a plurality of rule-based prompts, the plurality of rule-based prompts including prompts corresponding to a statistical term of a plurality of statistical terms;
providing the plurality of rule-based prompts and the tabular data to a first machine learning model with instructions for the first machine learning model to generate a plurality of natural language summary statistics, the plurality of natural language summary statistics including at least one natural language statistical summary for each statistical term of the plurality of statistical terms;
generating a plurality of instruction-output (“IO”) pairs, the plurality of IO pairs including an IO pair for each of the plurality of natural language summary statistics; and
training the machine learning model using at least a portion of at least one or more of the tabular data, the plurality of natural language summary statistics, and the plurality of IO pairs, wherein the training includes training the machine learning model to generate natural language statistically supported alerts;
generating an alert that includes at least one statistical relationship to be monitored between at least a first dataset and a second dataset;
monitoring data from the first dataset and the second dataset to determine that the statistical relationship is satisfied; and
in response to determining that the statistical relationship is satisfied, providing a natural language response to the alert that includes at least an indication that the statistical relationship is satisfied.
|