“…In our study, we employ the superior volatility measures to forecasts realized volatility (RV), which has been recorded in the numerous literature ( Fang, Wang, Liu, & Song, 2018 ; Liu, Ma, & Zhang, 2019 ; Liu & Zhang, 2015 ; Ma, Wei, Liu, & Huang, 2018 ; Peng, Chen, Mei, & Diao, 2018 ; Pu, Chen, & Ma, 2016 ; Wang, Wei, Wu, & Yin, 2018 ; Wei, Wang, & Huang, 2010 ; Wen, Gong, & Cai, 2016 ; Wen, Zhao, Zhang, & Hu, 2019 ). Subsequently, we utilize the heterogeneous autoregressive (HAR) model of Corsi (2009) , which has become the workhorse of forecasting models for stock volatility due to its simple linear regression techniques and consistently superior forecasting performance ( Buncic & Gisler, 2016 ; Cubadda, Guardabascio, & Hecq, 2017 ; Liang et al, 2020 ; Liu et al, 2019 ; Ma et al, 2018 ; Pu et al, 2016 ; Qiu, Zhang, Xie, & Zhao, 2019 ; Wang, Ma, Wei, & Wu, 2016 ; Wen et al, 2016 ; Wen et al, 2019 ). More specifically, Aganin (2017) compare the GARCH, ARFIMA and HAR-RV models and the results show that HAR-RV model has superior performance than GARCH and ARFIMA models; moreover, Vortelinos (2017) compares the forecasting performance of nonlinear models (Principal Components Combining, neural networks and GARCH) and HAR-RV model, the result indicate the simple HAR model is the most accurate for seven US financial markets (spot equity, spot foreign exchange rates, exchange traded funds, equity index futures, US Treasury bonds futures, energy futures, and commodities options).…”