Font Size: a A A

The Research On Time Series Factor Model Applied In Bank Plate

Posted on:2017-03-30Degree:MasterType:Thesis
Country:ChinaCandidate:Y WuFull Text:PDF
GTID:2309330503966672Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
With the rapid development of national economy and the market economy, the bank plate plays an increasingly important role in the national economy, especially in the financial industry. As the leaders of China’s financial industry, the 16 listed companies of bank plate in Shanghai and Shenzhen stock markets play a decisive role in the capital market. Therefore, it is of great significance paying close attention to the trends in the bank plate and the ups and downs of listed bank companies for observing trends in the bank plate and the capital market.The many listed bank companies provide a wealth of information for the analysis of stock price index of the bank plate, which on the other hand increases the complexity when dealing with the problem at the same time. The approach in this paper is from a dimension-reduction point of view. The daily returns of 16 stocks in the period 5 January2015–31 December 2015(244 days in total) are analyzed. Those stocks were selected among those included in the listed bank companies and were traded every day during the period. The paper is organized as follows. Section 1 is an introduction, describing the background and purpose of the bank plate stock price index research, and the method and structural of this article. Section 2 describes the conventional factor analysis method. The model and the estimation methods of dynamic and static time series factor model are introduced in Section 3, explaining this new dimension-reduction method in detail, and focusing on the ratio-based estimator for the number of factors, which is the theoretical basis of this paper. Then a simulation follows. The analysis of a real data set is reported in Section 4, in which the model is applied in analyzing the bank plate, the conventional factor analysis first, followed by the new methods. Section 5 deals with the cases when the old and new methods lead to different results, for which the new method is preferred. The results show that he ratio-based estimator leads to 1, indicating one factor. Thus the common factor is used to analysis the changes of the bank stock price index.The new contributions of this paper include: Firstly, the latest achievement of the ratio-based estimator for the number of factors is introduced in this article from abroad, which will provide references for domestic research. Secondly, differently from the factor analysis for independent observations, the factors which drive the serial dependence of the original time series are searched for.
Keywords/Search Tags:factor analysis, time series, dynamic model, static model, the ratios of eigenvalues, bank plat
PDF Full Text Request
Related items