Using the Doc2Vec Algorithm to Detect Semantically Similar Jira Issues in the Process of Resolving Customer Requests

Abstract

The paper is devoted to research in the field of software maintenance automation. An approach to solving customer requests based on the use of the Doc2Vec algorithm is proposed. It consists of finding semantically related resolved requests, as well as identifying qualified software engineers in the Jira bug tracking system. The developed software tool implements the proposed approach and provides reports which help software engineers in solving unresolved customer requests. The experiment compares the automated approach to resolving customer requests with the manual one. The results show advantages of using the software tool in the maintenance process.