The Empirical Analysis Of Statistics About Life Insurance Premiums Based On Time Series Analysis Method

As one of the important life insurance kinds, life insurance has influenced greatly on the social and economic life. This article analysis the data from the January 1,2012 to December 31, 2014 by the method of one time series and predict the future data from January 1, 2015 to 10. And all the data comes from Qingdao branch of GuoHua life insurance co., LTD. Found on January 1, 2015 to 10, the relative error between the actual value and the predicted values of the largest at 22.74%, the smallest is 1.44%. And it can be seen on January 1, 2015 to 5, the relative error is within 8%, and in 2015 after January 5, the relative error is above 12%, the average relative error was 10.19%. Due to many factors influence the life insurance income, for example, the degree of population ageing, market rate, inflation, the development level of other risks. Among them, the development level of the level of inflation and other risk had a huge impact on the development of life insurance. Therefore, in order to better predict future life insurance premium income, and further introduce the representative indicators CPI inflation factors and national insurance premium income the time series of the two indicators through the cointegration test, Granger causality test, and then establish multivariate time series model, and on January 1, 2015 to 10 of guohua life insurance premium income to forecast daily, prediction results show that using multivariate time series forecast GuoHua life future monthly premium income, the average relative error was 5.81%. Therefore, the use of multivariate time series makes the average relative error is smaller.