US 12,339,807 B2
Method and system for automating analysis of log data files
Sasikala Sethu, Madurai (IN); and Srinivas T, Hyderabad (IN)
Assigned to HCL Technologies Limited, New Delhi (IN)
Filed by HCL Technologies Limited, New Delhi (IN)
Filed on Aug. 27, 2022, as Appl. No. 17/897,135.
Prior Publication US 2023/0072123 A1, Mar. 9, 2023
Int. Cl. G06F 16/16 (2019.01); G06N 5/02 (2023.01)
CPC G06F 16/162 (2019.01) [G06N 5/02 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A method for automating analysis of log data files, the method comprising:
pre-processing, by an Artificial Intelligence (AI) model, a log data file to generate a transformed log data file, wherein the log data file comprises a set of data files in one of a plurality of data formats;
analyzing, by the AI model, the transformed log data file to detect one or more of a plurality of anomalies in the transformed log data file, wherein the one or more of the plurality of anomalies is detected by identifying one of a new and abnormal behavior associated with the transformed log data file;
performing, by the AI model, a predictive analysis for the one or more of the plurality of anomalies detected in the transformed log data file, wherein the predictive analysis is performed based on a set of factors and a separate confidence score associated with each of the set of factors;
wherein performing the predictive analysis based on the set of factors and the confidence score associated with each of the set of factors comprises classifying each of a set of errors in one of the plurality of error categories, further comprising:
categorizing, by the AI model, at least one of the set of errors in one of a pre-defined error categories; and
generating, by the AI model, a new error category for at least one of the set of errors, wherein the new error category is generated upon identifying a new error from the set of errors, wherein the new error is absent in the pre-defined error categories; and
wherein performing the predictive analysis based on the set of factors and the separate confidence score associated with each of the set of factors further comprises:
identifying, by the AI model, a set of errors in the transformed log data file based on the one or more of the plurality of anomalies detected in the transformed log data file;
classifying, by the AI model, each of the set of errors corresponding to the transformed log data file in one of a plurality of error categories;
identifying, by the AI model, a root cause associated with each of the set of errors identified corresponding to the transformed log data file;
recommending, by the AI model, a set of possible solutions based on the identified root cause for each of the set of errors; and
generating, by the AI model, the separate confidence score for each of the set of errors, the root cause associated with each of the set of errors, and each of the set of possible solutions, p1 generating, by the AI model, a report based on the predictive analysis performed for each of the one or more of the plurality of anomalies;
receiving, by the AI model, a feedback from an end-user based on the report generated for each of the one or more of the plurality of anomalies, wherein the end-user provides the feedback based on a pre-configured threshold value; and
updating, by the AI model, a database based on the feedback from the end-user for each of the one or more of the plurality of anomalies.