Research On Intelligent Discovery Method Of Network Service Oriented To User Personalized Preference

With the development of electronic technology, electronic commerce, ubiquitousservice and service-oriented technology, researchers have made great progress in thefield of the renovated service industry and integration between modern services andinformation industry. How to accurately and quickly discover the services required bythe user and provide effective recommendation under complicated networksenvironment has become a great challenge.Currently, the study of service discovery technology is fledge，while manyproblems still need to be studied in network service intelligent discovery, suchas:(1)From the perspective of services providers, renovated service industry is achallenge task. Because of the expansion of services sacle and scope, diversificationand innovative of service categories (includes Web Services, Cloud Services, Witkey,Ubiquitous Services et al.) and the realization of electronic upgrade and conversionfor traditional services, a problem need to be solved is services providers how torealize intelligent services discovery, selection and decision making. And then toimprove user services experience, realize the long-tail effect, achieve considerabledevelopment of services industry and maximize the enterprise value.(2)From theperspective of services users, the dual problem that services users faces the increasingof services information and the diversification of users services demands, the existingservices recommendation method can’t fulfill users satisfaction, and before usersdecision making, there are lot of works like services searching, information analysisand services attributes comparision. So that the existing services discovery methodsare with low efficiency.(3)Meanwhile, massive user services personalizedrequirements increase the complexity of intelligent services discovery andrecommendation.To solve those problems above, this dissertation seeks to use the long tail theory,network users behavior theory, pervasive computing technology, services discoverytheory, services composition theory, personalized recommendation technology andother related knowledge and methods, and it puts forward a PSO based intelligentservice discovery method oriented to user personalized preference. First, it studies intelligent services discovery and composition methods and user services relatedbehaviours, and presents a network service intelligent discovery framework orientedto user personalized preference. On that basis, it extracts user services preferenceinformation for user services history information and models an intelligent serviceuser model. Then, it studies the personalized service ontology initialization andoptimization method based on user interest degree and service popularity degree. Atlast, it puts forward a PSO based intelligent service discovery method as well asanalyzes with actual cases.The main contributions of this thesis are as follows:(1) It builds an intelligent service-user model. Facing problems that existingWeb service discovery methods are just focus on services research points: userservices function request and services quality demands, but ignore the vital influencepower of user services history information for user services selection and decision.This paper builds a service-user model based on the following questions: how to saveusers personalized information; how to describe users personalized information andhow mine users personalized service preference, and designs corresponding statisticalstandards of different user service-related activities, aims to solve problems of datasparseness and cold start through cross-correlation and social recommendation, solvesthe problem of the low utilization ratio of user personalized information throughmining users patterns of behaviours,then carry out their good business value andoptimize performance of service discovery.(2) It builds a service ontology model. Based on the contribution of traditionalWeb service ontology model in which services knowledge is provided by Webservices provider, this paper analyzes changes of users’ service-related activities,presents a personalized network service ontology initialization and optimizationmethod based on user service preference and demands. It also does some judgementsfor service node in PSO based on changes of network service user interesting degreeand network service popularity degree, and then takes the special additional steps,like adding operation and deleting operation. The special additional steps ensure thedynamic optimization of personalized service ontology. Network services ontologycan improve the efficiency and accuracy of personalized services discovery, raiseusers loyalty based on good users experience, and promote the achievements ofmaximization of enterprises value. (3) It designs an intelligent service discovery method under complicated servicesnetwork environment. In light of the customers’ needs and preferences for servicesdiscovery, it integrate service user model and personalized service ontology forintelligent services discovery. In order to meet users’ services requirements effectively,it researches the matching and mapping between user services requirements andcandidate services and designs personalized service ontology based service discoverymethod. This method contains researches of service name similarity computing,service attributes matix modeling and service relationship mining, realizes the theutilization of user services personalized preference, improves the efficiency of servicediscovery. It is also very significant to strengthen the Chinese modern serviceindustry by creating new patterns of network services.