Method and system for finding precursor subsequences in time series

摘要

A system and method for identifying patterns in time series data representing machine operation. The memory stores and provides a set of example training data generated by the machine sensors, each example training data representing a machine operation over a period of time ending with a machine failure. The processor repeatedly divides each training data example into a normal region and an abnormal region, identifies a prediction pattern that is absent in the normal region and exists only once in each abnormal region, and determines the length of the abnormal region Configured as follows. The predicted pattern is output via an output interface in communication with the processor, or the predicted pattern is stored in memory, and the predicted pattern is a predictive estimate of an imminent failure and assists in machine management.