Analysis And Prediction Of Stock Market Volatility Based On Xueqiu

In recent years, applying technologies of social networks to financial quantitative analysis, especially for predicting the volatility of stock market,is becoming a popular research direction in the area of data mining.Most existing works focus on traditional public social networks (e.g.Facebook and Twitter), or some other specific stocks. Comparatively, little is known about the emerging investor social networks or the full amount of stock market.In this paper, we study on one of the representative emerging investor social networks, Xueqiu. Based on the large-scale dataset collected from Xueqiu, we analyze the feature of the dataset, and identify the relationship between the dataset and the stock market from multi-perspective. Finally,we extract a series of quantitative features from the dataset, and build a set of stock market volatility prediction model by applying data mining techniques to the analysis results. Aiming at the full amount data of Chinese A-share, we conduct an empirical study on the effectiveness of using the dataset collected from Xueqiu to predict the volatility of stock market. In addition, we implement a practical distributed data crawler system to support the data collection. Our research has originality in practice and advances existing theoretical methodologies. We propose a series of new theoretical methodologies with good scalability and practical value.