Measuring and Forecasting Volatility in Chinese Stock Market Using HAR-CJ-M Model
Figure 3
(a) Comparison of the out-of-sample forecasting performance of the HAR-ARV, HAR-CJ, and HAR-CJ-M model (1 day). ARV represents the true volatility; HAR-ARV, HAR-CJ and HAR-CJ-M represent the forecast volatility of the HAR-ARV, HAR-CJ, and HAR-CJ-M models, respectively. (b) Comparison of the out-of-sample forecasting performance of the HAR-ARV, HAR-CJ, and HAR-CJ-M model (1 week). ARV represents the true volatility; HAR-ARV, HAR-CJ, and HAR-CJ-M represent the forecast volatility of the HAR-ARV, HAR-CJ, and HAR-CJ-M models, respectively. (c) Comparison of the out-of-sample forecasting performance of the HAR-ARV, HAR-CJ, and HAR-CJ-M model (1 month). ARV represents the true volatility; HAR-ARV, HAR-CJ, and HAR-CJ-M represent the forecast volatility of the HAR-ARV, HAR-CJ, and HAR-CJ-M models, respectively.