Regression parameters prediction in data set with outliers using neural network

Abstract

Popular regression techniques often suffer at the presence of data outliers. The different methods have proposed to make smaller the effect of the outlier on the parameter estimates. In this study, an algorithm has been addressed based on Adaptive network based fuzzy inference system to define the unknown parameters of regression model where dependent variable has outlier. So, three numerical examples are solved to test the activity of the proposed algorithm in regression model estimation. Also, the obtained results from the different methods, such as linear programming (LP) and fuzzy weights with linear programming (FWLP) are compared together. The results show that the proposed method is not to be affected the outliers in the solving process.