多因素时间序列法的研究

该文分析现有的单因素时间序列法在实际应用中的不足之处，提出采用多无线性回归原理综合处理的多因素时间序列法。它首先利用单因素时间序列法对各项影响事物发展趋势的因素进行预测；然后利用多无线性回归法将各种因素综合起来，以求得事物的总发展趋势。利用股票数据对单因素法和多因素法两者进行比较，可以明显看出多因素法较之单因素法在预测精度上的优势。

The Investigation of Multi-factor Time Series Method

The thesis analyses the deficiency of practical use of single factor of time series,and gives the method of multi factor of time series by use of principle of multi element linear regression.First use the method of single factor of time series to forecase factors that affect developing trend in things,then use the method of multi factor of time series to synthesize factors and to acquire general developing trend in things.And use the two methods to the model of stock price analysis.Compare the two results and find that the precision of multi factor time series method was higher than that of single factor method.