Software architecture design and development of multi-layer highly modular platform using intelligent components for dynamic big data analytics

摘要

This paper presents the design, development, and implementation of multi-Layer modular dynamic data analytics software platform using intelligent components. The reference software architecture proposed in this paper has been used to build the software robot called YaMasuta. YaMasuta is a software application that continuously analyzes massive financial data to produce investment decisions to aid human analysts. The objective is to investigate and test an optimal and efficient design of big data analytic software platform and designated as a reference software architecture platform for dynamic big data analytics' applications in different fields. The software platform of dynamic big data analytics will be presented at seven layers: The data acquisition layer, the data filtering layer, the data semantic layer, the intelligence layer, the model layer, the controller layer, and the dynamic presentation layer. Moreover, in this paper, we will present our chained asynchronous intelligent software components structure. The software robot and all of the intelligent software components described in this paper have been designed, developed, and deployed online at yamasuta.com for financial big data analytics and at yamasutaanalytics.apphb.com for medical morbidity analytics. Currently, the software robot YaMasuta for investment decision analytics covers financial data analytics of over 80,000 financial instruments (Stocks, FOREX, and Futures) in the largest 57 financial exchanges in 33 countries.