It(?) Semimartingale Test Of Asset Price Process In Chinese Stock Market

Statistical analysis of high-frequency data is one of the important contents of financial statistics research.In previous studies,people often assume that the volatility of asset price is characterized by semimartingale process.In particular,It(?) semimartingales are often preferred by people due to its excellent mathematical properties.Recently,some scholars studied the It(?)semimartingales hypothesis of asset prices,proposed the statistic to test It(?) semimartingales and tested the It(?) semimartingales hypothesis of the model based on the assumption of high frequency data.In this thesis,we use the method of convolution kernel smoothing to improve the statistic for testing the It(?) semimartingales,and obtain a new statistic by using the weighted p power variation,and indirectly verify the asymptotic normality of the improved statistic through Monte Carlo simulation.At the same time,using the statistics proposed by this thesis,some representative domestic price indices and stock price data were selected for empirical analysis,and the effects of different sampling periods and the degree of price volatility on the test results were investigated.The conclusions show that whether the price process is It(?) semimartingales is affected by the degree of price volatility and the length of the sampling period.