System and method for determining reasons for anomalies using cross entropy ranking of textual items

A system comprising:

• a processing device;
• and
• a memory device in which instructions executable by the processing device are stored for causing the processor to:
• detect an anomaly in a time series that is used to track metrices in textual data;
• identify a time step ai(t) associated with the detected anomaly;
• collect relevant textual data from a time window twindow associated with the detected anomaly;
• train a language model on the relevant textual data;
• for every textual item at time step ai(t), calculate a cross-entropy value according to the language model;
• and
• generate a set of textual items having a cross-entropy value greater that a predetermined value.

A non-transitory computer-readable storage medium having program code that is executable by a processor to cause a computing device to perform operations, the operations comprising:

• detecting an anomaly in a time series that is used to track metrices in textual data;
• identifying a time step ai(t) associated with the detected anomaly;
• collecting relevant textual data from a time window twindow associated with the detected anomaly;
• training a language model on the relevant textual data;
• for every textual item at time step ai(t), calculating a cross-entropy value according to the language model;
• and
• generating a set of textual items having a cross-entropy value greater that a predetermined value.