An Adaptive Neuro-Fuzzy Inference System (ANFIS) Method for Recommending Employee Mutations in a Manufacturing Company

Abstract

This research was conducted at Indonesia's leading rigid plastic and packaging company with a strong market position. At present, the company has expanded and opened branches in 6 countries in the Southeast Asia region. The problem occurred by the company is the opening of a new branch in Vietnam that requires several employees. The company issued a policy to select employees from companies located in Indonesia to meet the needs of employees in Vietnam. This study aims to facilitate the HRD in selecting employees who will be transferred. The total number of employees who meet the initial criteria for mutation are 666 employees. In this study, a recommendation system was developed that applies the Adaptive Neuro-Fuzzy Inference System (ANFIS) method. This method is a combination of Fuzzy Logic Method and Artificial Neural Network (ANN) Method. The criteria used in this study are employee attendance, duration of work of employees, employee psychology, employee job class, and employee performance. The results of this study indicate that ANFIS modeling with the hybrid algorithm along with the generated bell membership function can produce the best level of accuracy with an average Root Mean Square Error (RMSE) of 0.095561. We also developed a decision support system that applies the ANFIS method. Based on the quality testing of the DSS using the Software Quality Assurance (SQA), it obtained a quality value of 82.6%, which is in the range of good grades.