Predicting newspaper sales by using data mining techniques

Abstract

This study focuses on the newspaper sales prediction problem. This is a critical shortcoming especially for newspaper companies, which always try to reduce the number of sell outs and the number of return in the delivered newspapers. This is crucial for decreasing the cost and increasing the profit. It is basically a time-series problem, so it is attempted to be solved with the Holt-Winter's, Linear Regression and Fuzzy C-Means(FCM)+Adaptive Neuro Fuzzy inference system(ANFIS) methods. The relationship between the inputs and the outputs are investigated and calculated by these three methods. The results are presented in the findings section, where they are also compared. The best method for each outlet is selected according to forecast evaluation measures. © 2021 Elsevier B.V., All rights reserved.