Application of Kernel Density Estimation in Stock Trading Volume Distribution

Stock trading volume is an important indicator reflecting market activity and investor behavior. It is of great significance to study its distribution characteristics and dynamic changes for understanding market behavior. This study uses kernel density estimation to analyze the probability distribution characteristics of trading volume in A-share market, and combines long-term short-term memory network model to model the time dynamic change of trading volume. The results show that the stock trading volume shows significant right-skewed and fat-tailed characteristics, which is suitable for the generalized inverse gamma distribution. The long-term short-term memory network model can effectively capture the time trend of trading volume, and the time slice and rolling window analysis of kernel density estimation reveal the high activity of the market in the opening and closing periods. This study provides data support and method reference for market volatility analysis and transaction risk management.