High Prediction Accuracy and Low Error for ERP User Satisfaction by Hybrid of ANFIS and KNN Classification

The incoming era is becoming more friendly and dependent on Information Technology. Enterprise Resource Planning (ERP) Systems are one of the most widely used latest examples of Information Systems (IS) technology. They provide a single window system to the organizations by integrating the whole functions of them. Today, all enterprises are rapidly adopted ERP systems. But, their adoption and implementation is not being without any problem. The implementation process of ERP is also a very challenging, time consuming and costly task. Therefore, instead of many efforts if the implementation process is failed. Then it will be a big failure for the organization. Hence, to overcome this failure and increase the success rate of ERP projects we need to develop a robust, reliable and accurate predictor. This will help us to redirect the projects far better in advance. The success of ERP systems depends on many factors. US is one of the important factor among them. Hence, we develop an efficient predictor of US using hybrid of ANFIS and KNN. We were used this method first time in literature related to prediction of US in ERP. The Hybrid method increases the prediction accuracy more comparatively than previous reported techniques ANN, ANFIS and KNN. The RMSE using Hybrid method is 0.167629 and for KNN, ANFIS and ANN is 0.5, 0.486185, and 0.590329 respectively.