New Dataset for Forecasting Realized Volatility: Is the Tokyo Stock Exchange Co-Location Dataset Helpful for Expansion of the Heterogeneous Autoregressive Model in the Japanese Stock Market?

This study analyzes the importance of Tokyo Stock Exchange Co-Location dataset (TSE dataset) to forecast realized volatility (RV) stock price index futures. The heterogeneous autoregressive (HAR) model is a popular linear regression used RV. expands HAR using TSE dataset, full-board and market volume based on random forest method, which machine learning algorithm nonlinear model. new dataset. only information that shows transaction status high-frequency traders. In contrast, buying selling dominance. as proxy for liquidity recognized important in finance. To best our knowledge, this first use co-location experimental results show yields higher out-of-sample accuracy RV than Moreover, we find has become more recent years, along with increasing trading.