Abstract
PCA has been widely used in many fields to detect dominant principle components, but it ignores the information embedded in the remaining components. As a supplement to PCA, we propose the General Component Analysis (GCA). The inverse participation ratios (IPRs) are used to identify the global components (GCs) and localized components (LCs). The mean values of the IPRs derived from the shuffled data are taken as the natural threshold, which is exquisite and novel. In this paper, the Chinese corporate bond market is analyzed as an example. We propose a novel network method to divide time periods based on micro data, which performs better in capturing the time points when the market state switches. As a result, two periods have been obtained. There are two GCs in both periods, which are influenced by terms to maturity and ratings. Besides, there are 382 LCs in Period 1 and 166 LCs in Period 2. In the LC portfolios there are two interesting bond collections which are helpful to understand the thoughts of the investors. One is the supper AAA bond collection which is believed to have implicit governmental guarantees by the investors, and the other is the overcapacity industrial bond collection which is influenced by the supply-side reform led by the Chinese government. GCA is expected to be applied to other complex systems.
Highlights
Principal component analysis (PCA) is one of the most established methods used in the fields of science and economics [1, 2]
Through the PCA method, the components with the largest eigenvalues are selected as the dominant PCs, which contribute to a large fraction of variances
We are inspired to propose General Component Analysis (GCA) to extract the systematic information from all of the components. Both of the global components (GCs) and localized components (LCs) are identified according to their localization property
Summary
Principal component analysis (PCA) is one of the most established methods used in the fields of science and economics [1, 2]. It is the case especially in the corporate bond market, where the remaining components reveal important information, but only the top-ranked PCs have been concerned [7, 12, 13]. The second step is to analyze the information embedded in the GCs. The third step is to construct the portfolio of the LCs and use the complex network approach to detect its correlation structure.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
