User Satisfaction Prediction in ERP using KNN Classifier for high Prediction Accuracy

Abstract—ERP (Enterprise Resource Planning) systems are widely used in organizations; because, ERP provides a single platform to manage all the processes and functions of organizations. This single platform improves their productivity, business performance, decision making capabilities and efficiency. However, to achieve a proper level of ERP success depends on various factors e.g. organization, technology, environment and User Satisfaction etc. ‘User Satisfaction’ (US) is most important factor to make ERP successful. US refer the user’s comfort and acceptability of ERP system during the use and interaction with the ERP system. This paper deploys the conceptual model for US prediction by considering Human, Technological and Organizational factors as predictors. In this report, we proposed K-Nearest Neighbor (KNN) Classification method first time in literature to predict the US and we compare it with ANFIS and ANN. We calculated average error for all test cases and demonstrate that KNN gives high predication accuracy in most of the cases and low average error (0.25) in comparison ANFIS (0.3378) and ANN (0.6053) methods. So our approach is novel and using KNN, prediction accuracy can be further improved for US to make successful ERP.