Indian Journal of Science and Technology

Abstract

Objectives: Systems managed by ITSM include incident management process. Considering the characteristics of incidents, much professional manpower and time should be needed. The purpose of this study is to suggest AIPS (Automatic Incident Process System) for simple and repeated incidents among those registered on ITSM to improve existing inefficient shortcomings. Methods/Statistical Analysis: AIPS uses NBC (Naive Bayesian Classification) to automatically classify incidents registered on ITSM to simple class or complex class. With NBC, AIPS learns accumulated processed data and distinguishes simple and repeated incidents. In addition, it calculates the frequency of main solutions for the distinguished incidents, and then transmits the solution with the highest frequency to requesters through e-mail. Findings: 52,735 incidents that had been processed for six months were analyzed, and 9,238 of them were confirmed to be simple and repeated incidents, which account for 17.51%. AIPS and human tried to analyze and classify the same incident. As a result, AIPS using NBC classified 7111 (76.97%) incidents to simple and repeated incidents, which were smaller than 8,603 (93.12%) classified by human. A proper solution was selected according to the solution frequency of the classified incidents. The selected solution was sent through e-mail to grasp satisfaction. More than 87% of requesters answered “Satisfied (over score 80)”. The main reason of satisfaction was “Quick processing (89%)”. Simple incidents have a merit that they can be quickly resolved if managed in time. However, it is difficult to search simple incidents among a huge number of incidents and process them. AIPS suggested by this study was used, the working hours of professional manpower could be shortened by 77%, reducing the degree of workload, which raised concentration on processing of complex incidents. This will improve the overall working efficiency. Improvements/Applications: Owing to automatic processing of simple incidents, not only working efficiency but also requesters’ satisfaction will be improved.
Keywords: Automatic Processing, Incident Process, ITSM, Naive Bayesian Classification