COMPARISON OF PERFORMANCES OF CLASSIFICATION METHODS IN ESTIMATION OF STOCK EXCHANGE INDEX IN BIST
BİST ENDEKS HAREKET YÖNÜNÜN TAHMİNİNDE SINIFLANDIRMA YÖNTEMLERİNİN PERFORMANSLARININ KARŞILAŞTIRILMASI

Author : İsmail KARAKULLE -- Fatih ECER
Number of pages : 514-524

Abstract

Recently, the method of estimating the price or the yield of the stock index has been used more frequently than the statistical methods. The aim of the study is to compare the performance of various classification methods in the estimation of BIST Bank index movement direction. While 10 technical indicators calculated in the BİST were used as input variables for the analysis methods in the study, the next day's stock market index was used as the closing value output variable. According to findings, the correct classification performances of artificial neural networks (ANN), support vector machines (SVM), logistic regression (LR) and linear discriminant analysis (LDA) models are 81.74%, 60.87%, 76.70%, 76.87% respectively. The results show that machine learning methods are effective methods that can be used in estimating the index motion direction.

Keywords

Classification, Machine Learning, ANN, SVM, BIST Bank.

Read: 705

Download: 232