Research On Personalized Information Retrieval Based On Context-aware

With the explosive growth of resources on Web, a lot of problems have occurred such as information overload. Traditional retrieval hasn't resolved similar problems. Contextual Information Retrieval has been focused in recent years. In the process of information retrieval, users and the information which is needed by users are in the contexts of their own. Contextual Information Retrieval is to rational use of these contexts such as user context page context link context and equipment context.Contexts combine with information retrieval can improve the retrieval performance. This paper research user interests modeling which is used by contexts and further achieve the personalization information retrieval based on contexts.The chief contents of the paper are as follows:(1)The paper contrast traditional information retrieval with the contextual information retrieval and show the latter have an advantage over the former. Contextual information retrieval centers on computer and learn about the contexts to adjust the user interests modeling. So it can improve the retrieval performance.(2)The paper studies the contexts of web which can influence user access behavior.Page context and user context are thought to be more influence user access behavior. It analyzes the behave context of browse in user context and uses the browsing speed in the experiment of user interest model.(3)The paper gives the hierarchical VSM based on contexts to express user interests. The basic idea is that firstly extracting key words in the documents to express user interests. Secondly using the tree structure separate and group the user interests. Thirdly using the VSM describes the user interests.(4)Because of the advantage of XML, the paper uses the XML to save user interests. The inverse index improves the speed of retrieval.