| With the rapid development of Internet,based on B/S architecture's development technology tends to be mature.The application which is web-oriented has been gaining popularity and attention more and more.This thesis embarks from opening business processing model,combines the requirements of YNSISS(scientific instrument shared service platform of Yunnan)and makes an intensive study of the key technology of business processing model,instant information interaction and intelligent recommendation.An opening business processing model has been presented,a web-oriented general short message assisted service platform has been built and an intelligent recommendation engine based on case-based reasoning has been implemented.The traditional centralized business processing model and usual opening business processing model are analyzed.The key problem must be resolved in the transduction of centralized control into multi-grades decentralized control is investigated,such as grading authorization policies,data isolation,applying and accepting flow,instant information interaction based on isomerous communication,opening identify and certificate authentication,and so on.Based on this,the personalized service guiding model will be further studied,which can guide user to deal with correlative business automatically according to user's rose,grades and current business state.A general opening business processing model is presented,which has been used successfully in YNSISS.The definition and design requirement of general short message assisted service platform are anatomised in detail,the communication model of short message platform in china is investigated,the multi-layer architecture of communication model independence is designed,the core working flow of short message assisted is designed,the flow of short message sending and receiving,short message pool,scheduling policy,distributing policy and billing are studied,a web-oriented general short message assisted service platform is built,which has been integrated into YNSISS Successfully and has resolved the problem of instant information interaction in the opening business processing model.The design goal of intelligent recommendation engine is presented,multi-layer architecture that supports multi-datasource is designed,the core working flow of intelligent recommendation engine is analyzed,the technology based on case-based reasoning is studied,the core problem of case base representing,feature extraction, weight scoring of case feature,similarity computing,multi-layer index mechanism based on feature,case adaptation,case persistence are studied in depth.The technology of case-based reasoning is fused into intelligent recommendation and a general intelligent recommendation engine can be built,which uses the develop method based on component, according to configuration.The function of instrument recommendation has been implemented in YNSISS with this engine.The problem of intelligent recommendation and instant information interaction have been resolved in the opening business processing model according to the construction of web-oriented general short message assisted service platform and intelligent recommendation engine. |