Patent · US Active

Machine learning-based recommendation system for root cause analysis of service issues

US11176464B1 · kind B1 · utility

10Cited by
10References
20Claims
0Family size

Assignee

Inventors

Key dates

Filing dateApr 25, 2017
Grant dateNov 16, 2021
Priority date
Expiry dateApr 13, 2040

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06N5/022
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

A machine learning-based recommendation system is provided for root cause analysis of service issues. An illustrative method of a machine learning system comprises extracting features from service issue investigation log sets corresponding to previously considered service issues; generating representations for the service issue investigation log sets based on the corresponding extracted features; and storing the representations in a knowledge base. In conjunction with obtaining an additional service issue investigation log set, the illustrative method generates a representation of the additional service issue investigation log set; identifies representations previously stored in the knowledge base as candidate service issues related to the additional service issue based on pairwise probabilities indicating whether the additional service issue is related to at least a subset of the previously considered service issues; and presents information characterizing the service issue investigation log sets corresponding to respective ones of the identified representations.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.