The Research And Implementation Of Exploratory Search Based On Personal-model Within The Cloud Computing Framework

With development of information industry especially the Internet andMobile Internet, information data is rapidly growth; the day of lack informationhas gone. To some extent, people can get any information from the Internet. Butin the face of such tremendous information resources, people need an effectiveway to find the information they really interesting. So how to enhance the userexperience during the process of search yet to be resolved.Because of the personalization is not enough during the process ofinformation retrieve, So In our paper, on the basis of the traditional searchengine. We study of information classification, information clustering andSimilar Link Network; use personalized model to support visual explorationsearch. Nowadays, as the touch screen appeared. It well suited for our research.At last we will develop a new type visual exploratory search prototype whichbased on the user model.The content of our study as follow:1) Analyzed the existing search engines. Development a Meta-searchengine which integration of some general search engines and microbloginterface. At the same time, Developed an crawler software which based onopen source WebCrawler and Solr, use this crawler crawling information fromWeb。2) The information from the Meta-search is redundancy, the semantics arescattered, and it is difficult to find the relationship from the information. So we study content extraction, web pages redupliction removing, word cut and so on,use the improved K-means to classify and cluster the information. According tothe semantic similarity generate Similar Link Network. From discard theredundance and select the essential to reduce the size of the information to builda more concise and reliable semantic similarity figure. Lay the foundation forexploration personalized search.3) Nowdays, the existing search engine is not meet the require ofpersonalization. So on the basis of similar semantic link network we designed auser model. Use the model segmentation and positioning the similar semanticlink network, verify the search results to meet the needs of the userspersonalized search.4) Based on the information classification and clustering, similar semanticlink networks, user model; tracking the process of user’s searching activity,Combine the touch screen and Use graph to represent the process of exploratorysearch,research an new RF mechanism to support exploratory search.5) Based on Hadoop and HBase to build a cloud platform. On the platformwe use the model of personalized exploratory search system prototype todevelopment two practical application–“public opinion analysis system” and“wisdom district” which based on Android.