A Neural Network Based Model for Project Risk and Talent Management

Abstract

Objective of this research was to examine the effectiveness of using neural and fuzzy systems in the areas such as job recruitment, prediction of project success/failures and decision making process of employee performance appraisals related to any company/industry. Since these activities are common in many areas, the domain of the research was confined to the software industry. Through a thorough survey, most relevant parameters (together with their level of relevance) that are used in evaluating the suitability of candidates in recruitment etc. for various designations were gathered and fuzzy logic, neural network based approach was used for training and testing. Performance appraisals were also implemented using a neural network which was able to pinpoint the employee’s appraisal to a great deal of accuracy. Further, the prediction of project success and failure was also implemented similarly. Since many linguistic terms are used in these activities, a fuzzy input/output interface was accommodated to the system. The evaluation of the proposed system with the support from the industry experts shows a very high level of accuracy.