Estimating fuzzy nonparametric regression model based on some smoothing methods with practical application

Abstract

In this research, fuzzy nonparametric estimation methods based on some smoothing methods were applied to real data on the Iraq Stock Exchange, as the trading activity of Baghdad Soft Drinks Company was studied for the year (2016) (for the period from 1/1/2016 to 12/31/2016). A sample of (148) observations was obtained in order to build a model for the relationship between stock prices (lower, higher, and average) and trading volume. And by comparing the results of the (G.O.F) criterion, the three smoothing methods, and the kernel functions adopted and proposed by the researchers, it was noted that the lowest value of this criterion was for the nearest neighbor method (K-NN) and the kernel functions (Gaussian).