Systems and methods for enhanced machine learning using hierarchical prediction and compound thresholds

Abstract

A computer device is programmed to receive a plurality of real-time datasets from one or more sensors associated with a tool to be analyzed, calibrate the plurality of real-time datasets, generate a time slide window for each real-time dataset of the plurality of real-time datasets, generate a random probability distribution curve, compare the random probability distribution curve to each time slide window to determine if the time slide window includes anomaly data, and generate prediction results based on the comparison.

Images