Analysis of investment risk in companies listed on the Iran stock exchange ‎using ‎a ‎boosting algorithm

Zamani Mehreyan, Sedigheh; Eghbalifatr, Naser

doi:10.22034/jr_iss.2024.732974

Analysis of investment risk in companies listed on the Iran stock exchange ‎using ‎a ‎boosting algorithm

Document Type : Original Article

Authors

Sedigheh Zamani Mehreyan

Naser Eghbalifatr

Department of Statistics, Imam Khomeini International University, Qazvin, Iran.

10.22034/jr_iss.2024.732974

Abstract

In recent years, stock price prediction and market return models have attracted significant research attention. Traditional statistical forecasting models, such as autoregressive (AR), ‎autoregressive moving average (ARMA) and generalized autoregressive conditional heteroskedasticity (GARCH) models, have been widely used. Advances in computer processing power have facilitated the introduction of new algorithms, including machine learning techniques, for predicting financial data. Consequently, this study analyzes investment risk in companies listed on the Iran Stock Exchange by examining data on their trading volume. An optimal vector autoregressive (VAR) model is fitted using a boosting algorithm. The performance of this boosting-based model is compared to that of models fitted using conventional statistical methods. The results demonstrate the superior performance of the boosting algorithm.

Keywords

Boosting algorithm

Trading volume

Vector autoregressive (VAR) model

Generalized autoregressive conditional heteroskedasticity (GARCH) model

Subjects