Ultra-high-frequency data is tick-by-tick data, which has more information than low frequency data. Traditional GARCH model is suitable for the data with same intervals. Ultra-high-frequency data is tick-by-tick data with different time intervals. As a result, we can model the volatility of per interval of time. Engle puts forward UHF-GARCH model for modeling ultra-high-frequency data. This paper does empirical research of UHF-GARCH type models, which include UHF-GARCH-N model, UHF-GARCH-t model, UHF-TGARCH-N model, UHF-TGARCH-t model and UHF-EGARCH model, by using ultrahigh-frequency data of Shanghai stock market.
