基于Copula模型的数据分析平台的实现

Development of the Data Analysis Platform Based on Copula Model

以开源的企业级JBoss Seam为集合框架,采用了该框架中的Richfaces作为人机交互界面,EJB3 Session Bean作为业务逻辑层,JPA持久化到Oracle数据库,设计了基于Copula模型的数据分析平台.该平台以开源的R语言作为计算引擎实现了Copula算法,对上证指数和深成指数5年来的数据进行了相关性分析,给出了时间序列图,并可以很方便地得到Copula参数值和密度图,为股指的相关性分析提供决策支持.

The data analysis platform was developed based on Copula model. JBoss Seam was adopted as collections framework, which is open source and enterprise class. In the framework, Richfaces were used as the human--machine interactive interface, the EJB3 SessionBean was employed as the business logic layer, the J PA （Java Persistence API） stored objects to Oracle database. The platform used the R language as engine, based on the 5 years of data, correlation analysis between Shanghai Composite Index and Shenzhen Component Index was carried out, and time series chart was plotted. Copula parameters and the density maps also can be easily calculated by our platform, so it provides a useful decision support for the correlation analysis of the stock index.