| The survival analysis has been continually promoted in the field of epidemiology, actuarial science and so on, but rarely used in the financial industry. This paper devotes to the applications of the Cox model to the stock market, studying the Shanghai Stock Exchange(SSE 100) Index, in order to obtain the covariates that have important influ-ence on the stock and predict the probability that a stock keeps to survive over some time.Firstly,13 finance indicators having important influence on the stocks’ price are collected from the Da-Zhi-Hui software. Observation period is from January 2014 to July 2014. The survival time and survival state, two sets of datas about the stocks, are obtained by defining the time of failure.Then, based on the datas of each quarter. The coefficients between 13 finance indicators are explored to indicate the interdependency. Using the minimization of LASSO-type partial likelihood with penalty function λΣ|βj|by coordinate descen-t algorithm and cross-validation method for turning parameter A. So the important covariates are obtained.Finally, The significance of the covariates’parameters is tested by likelihood ratio test under the predetermined significance level α=0.05. The results show that using the important covariates which are obtained by variable selection to get survival curve is reasonable. Given a stock’ important covariates that are collected at the beginning of the quarter, the survival curve of the quarter are exploited which can predict the probability that a stock keeps to survive after a time period. The results can provide useful information for investors, with a strong practicability. |
PDF Full Text Request