Research And Application Of Multivariate Statistical Analysis Based On Data Warehouse

Statistical analysis software package has been studied and applied since 70s of last century.Nowadays international famous statistical packages have SPSS(Statistical Package for the Social Science) and SAS(Statistic Analysis System) and so on. Although the domestic research and application in this area start early,the systematic research and realizing the general statistical software packages of the current combination of network and multi-system computing environments are rarely reported.With the network economy and information technology development,a growing number of enterprises and institutions are not satisfied only by automation of business management through the application of computer systems,business has started to pay attention to the data analysis in the hope found in the data base of hide Internal relations,potential and trends in the law in order to effectively support the enterprises and institutions on the management,production and research and development,restructuring,operating and marketing decisions.The author of this paper participated in the development of the General Statistical Package A prototype system for the current network and multi-system computing environments,on the basis of participation in the research of mainstream of current statistical analysis software packages,as well as data warehouse technology based on multivariate statistical analysis of the feasibility of software and related technical characteristics.The system is based on the realization of three-tier architecture model, extracting data automatically from source data in the production environment according to preset parameters and the analysis model,and loading the data into data warehouse after conversion,with cross-platform data mart applications at the same time,providing a wide range of description of the statistics types and multivariate statistical analysis.The primary research work of the author of this paper participated in are the followings:1) Participate in design and realization of interactive statistical software package for the three-tier architecture;2) Research multivariate statistical model for interactive statistics software package,including the theory,methods,algorithms and human-computer interface;3) Based on data analysis model,participate in the research and realization of extracting,transforming and loading data into data warehouse from the production environment of Line Center Computer System and its data storage structure; 4) Define the structure of dimension table and fact table in star-dimensional data model to support OLAP data analysis on multi-dimensional model,according to the analysis and forecasting needs;5) Define XML format between layers of feedback and requests interface protocol;6) A full technical solution of time series applied on the forecast of passenger flow of rail transit has been given.Above work completed by the author of this paper is based on research and realization of "Automatic fare collection(AFC) operation management data analysis system of rail transit",which is checked and accepted by Shanghai Science and Technology Committee on July 12th,2007,and registered and awarded the certificate of science and technology achievement(Registered No.:9312007Y1168) on August 7th,2007,and the certificate of computer software copyright(Registered No.: 2007SR13214) by the National Copyright Bureau on August 30th,2007.