Key Technologies Research On Personalized Information Service Based On Ontology And Context-aware

With the rapid development of Internet and the level of electronic commerce informationimproving, amount of information people gain from them also has grown in a high speed.Thus, it results in “information overload” and “information lost”. Therefore, personalizedinformation service (PIS) has been gradually paid more attention, which could take thedifferences of users’ interests into account to provide different information service fordifferent users. However, traditional personalized information service has neglected contextsensitivity of individual requirement and semantic differences of users expressing their needs,which causes users’ requirement acquired is inaccurate and unreliable.In view of problems above in the present personalized information service, this paperputs forward the framework of PIS based on ontology and context-aware by using research ofkey technologies research on personalized information service in Spring Plan. And it focuseson user’s context ontology modeling and technologies of PIS base on context-aware. Theprimary works in this paper are presented as follows:(1) According to the current problems in PIS, this paper puts forward the framework ofPIS based on ontology and context-aware, and discusses work and process of the variousstages in detail.(2) By analyzing these advantages and disadvantages of methods of context modeling,this paper uses the context ontology modeling, and proposes LESSM context modelingmethod combining evaluation and the iterative algorithm in software engineering modelingbased on the original seven step method. In addition, it makes use of multi-granularityapproach to model user context; moreover, this thesis gives an example of some contextontology modeling about college students’ costume recommended system.(3) Using the forgetting theory in psychology for reference, I discover contexts of users’browsing have been forgotten to a certain degree. Consequently, non-linear gradual forgettingalgorithm has been presented to synthesize user’s historical context-information and currentcontext-information so as to get the user’s comprehensive context. It can diminish the weightof historical context gradually according to the context-information in that time.(4) This paper proposes contextual similarity recommendation algorithm. It can gain themost similar source context sequence through calculating the similarity of between current user’s comprehensive context and source context in which customers brought goods recordedin the database. And then we recommend the purchase commodity in these source contexts tocurrent user. We divide the context ontology tree into many children tree. According tocalculating the similarity of children tree with the weight sum, we obtain the similaritybetween the source context and target context. Meanwhile, during calculating the similarityof children tree, this paper allows for ontology semantic similarity, and adds the structuresimilarity degree algorithm on the basis of the old ontology similarity degree algorithm,which makes the result more accord with actual, and finally the experiment is verified.