Modeling and Forecasting Stock Market Volatility of CPEC Founding Countries: Using Nonlinear Time Series and Machine Learning Models

The highly sensitive, nonlinear, and unpredictable stock marketbehaviours are always challenging for researchers. Stock markets ofPakistan China, i.e., KSE-100 SSE-100, respectively, thetwo most attractive after the official announcementof CPEC. Thus, daily closing price of SSE-100 Stockreturns used to evaluate volatility forecast performance ofthe machine learning technique, GARCH family nonlinearregime-switching models. findings this study revealed that thestandard model is best-fitted based on Akaike’sInformation Criteria (AIC) Bayesian Information (BIC).Furthermore, LSTMmodel outperforms other models RMSE SSE-100. Incontrast, CGARCH theMarkov-regime-switchingmodelforKSE-100outperformsothermodelsbased MAE, MAPE, SMAPE evaluation criteria. It also revealedthat predictive power very closeto MRS model; therefore, LSTM can be asan alternative regime-switching marketvolatility. These will help national international investors,policy-makers, geographical economists, industrialists use thebestforecastmodeltomakebetterpoliciesandgaintremendousprofit.