On The Latent Factors Behind The Movement Of Real Estate Stock Price-Study Based On Independent Component Analysis(ICA)

As the section of real estate stock market has a major impact on the entire stock market and China's economy development, it is important to study the movement of the real estate stock price. There are a lot of factors that affect the movement of the real estate stock price, some of which are observable, and the others are latent. Find out the latent factors is rather challenging. At present, most researches focus on the observable factors with the traditional statistical methods (such as regression analysis and Granger cause-and-effect test), which, however, are incapable to find out the latent factors. In this paper, we apply a new analytical method called Independent Component Analysis (ICA) to study the latent factors behind the movement of the China's real estate stock price.It is now exploratory to apply ICA on financial time series. ICA is a method which is based on information theory (mainly on Neg-entropy and kurtosis) to find out the latent variables. The task is to transform the observed data x, using a linear static transformation W as s = Wx into maximally independent components s measured by some function of independence. ICA can effectively curb the Gaussian noise, thus it has a unique advantage in searching the latent factors.Applying ICA in the selected representative A-share real estate stocks, it is found that two major factors are driving the movement of China's real estate stock price: the big board and the Property financial policies. Besides, this paper use TAR model in the ICA to forecast the trend of the real estate stock price, which over performs the forecast with traditional forecast methods.