A Hybrid Model of Machine Learning Model and Econometrics’ Model to Predict Volatility of KSE-100 Index

Abstract

Purpose:

The purpose of this paper is to predict the volatility of the KSE-100 index using econometric and machine learning models. It also designs hybrid models for volatility forecasting by combining these two models in three different ways.

Methodology:

Estimations and forecasting are based on an econometric model GARCH (Generalized Auto Regressive Conditional Heteroscedasticity) and a machine learning model NNAR (Neural Network Auto-Regressive model). The hybrid models designed with GARCH and NNAR include GARCH-based NNAR, NNAR-based GARCH, and the linear combination of GARCH and NNAR.

Findings:

In a comparison of the forecasting results of the KSE-100 index over different periods, the least RMSE is found in a linear combination of NNAR and GARCH, followed by NNAR, GARCH, NNAR based GARCH, and GARCH based NNAR models.

Conclusion:

The study concludes that the hybrid model designed with a linear combination of GARCH and NNAR performs better among all the models in forecasting the volatility of the KSE-100 index.