Design And Realization Of Logistic Regression Analysis In 3-tiers Calculating Architecture

In the near several years past, the development of city rail transit already becomes the important approach to solve city public transportation. Managers of rail transit operation want to search for meaningful and valuable information to ensure that the rail transit system operation is safe, reliable and efficient from a large number of data. The project <> that this author participated in research is the first domestic data analysis application software package with statistical functions for general and special analysis purposes at present which is designed based on 3-tier calculating architecture pattern aiming at the characteristics of automatic fare collection (AFC) operation management and data sources of rail transit. It combined various statistical methods can be used to analyze AFC operation data of rail transit in order to find out the inner relationship and knowledges in the large amount of data, and to forecast the future tends in terms of the data accumulated in history for providing science data to decision of forward-looking and knowledge-based.According to the descriptions of 3-tier architecture pattern and its functions in each layer of AFC data analysis system, the communication transmission between the client and calculating server has been designed and realized in this paper by using of the regulation of short synchronized link with TCP/IP protocal.The Logistic regression model and its algorithm have been studied in this paper. Logistic regression model is the statistic analysis method in common use to analyze the response variables as independent classification variables. It is also the newer one in data mining methods, and has been extensively applied in the social sciences at home and abroad such as sociology, psychology, demography, political science, economics and public health. After integration processing of extracting, transforming and loading for rail transit AFC operation data, through analyzing, finding out that the logistic regression is suitable to forecast the terminal trouble affairs stored in line central computer system of rail transit AFC system in order to find out the primary affecting factors of the terminal troubles, to provide the data to make decision correctly for maintenance and the spare parts' supplying, and warranty of efficiency and reliability of the AFC system operating for managers.The ultimate research purpose in this paper is to put the logistic regression model in application of AFC operation management data analysis system of rail transit based on the 3-tier architecture pattern first through analyzing the characteristics of the rail transit AFC operation data, including the algorithm research and completing an example of business analysis in total solution.