| With the rapid development of computer technology,the flow of information is much faster and the amount of data hugely increases,which makes the traditional storage tools and computing tools no longer satisfy data’s growing scale.In this situation,Google put forward new theories about distributed file system,data processing frame on large clusters and distributed storage system for structured data.It opened a new era of big data and after which many tools of big data based on those theories came up.These tools work well with the growing amount of data.Event study is a research method in empirical finance.It measures whether an event has impacts on an underlying asset through quantitative methods.By reading papers of event study,fact is found that event study suffers from the poor ability of events collection.And many researches lack big data tools for storage,processing tools and a one-stop study platform.To solve those problems,this paper focuses on building a real-time event study system with full-scale data and computing engine of high efficiency.The main jobs are listed as follows:(1)Finish the collection and storage of company announcements data and stock data,and the announcements could be queried.Web spider is took advantage of during the collection,and this paper also use streaming middleware for collection and extraction.The collected data is effectively put into different places and get indexed for query.(2)Implement the calculation module under the big data computing framework.Make the design and implementation of forecast model and hypothesis testing model of event study,which improves the ability of data processing by parallel computing on clusters.(3)Complete the design of one-stop event study platform.Through the zonal design,including data collection zone,data storage zone,data process zone and user zone,this platform can offer many varieties of services.It saves much energy for users from data collection and process,then users can focus on event study exactly. |
PDF Full Text Request