Forecasting algorithm developed from the neuro - Fuzzy system

Abstract

© Copyright 2015 ACM. Forecasting is the process to estimate the future by analysis the past and present information. In data analysis the accurate forecast results are usually based on the reliable classification algorithm. This paper proposes the new Neuro-Fuzzy algorithm combined with the quadratic function. The classification performance of the propose algorithm is shown by the accuracy of 7 standard datasets from the UCI machine learning repository. The satisfactory classification accuracy is shown in the experiment. The proposed structure can classify by both direct calculation and using rule based classification. The classification structure after training process is easy to apply in any application.