A Study On Web Page Recommendation Algorithm Based On User Access Paths

With the development of Internet and its sufficient information, people became more and more to rely on Internet to search available information themselves, as a result, the search engine emerged. However, search engine failed to offer available information for their customer and failed to estimate information purpose without proper guidance to customer’s tendency. Therefore, a new intellect recommendation item for customer which based on customer’s searching action became a hot researching point. This paper mainly argues that when a customer searches something on the Internet, the item would give a webpage recommendation to probe the customer’s demand which bases on customers’before invitation webpage. Furthermore, the recommendation results could be used in the search engine to check out results to direct the main orientation and it could also be used to optimize the hyperlink hierarchy of web site.The main logic sequences of this paper could be represented like following steps:firstly based on internet customer’s invitation webpage and mining user’s paths, secondly the item will recommend less than8inter-navigation links for the customer; moreover these links reflect the most customers’information demand, finally the crosscut to offer customer satisfactory information is the final target. It is noted that this item belongs to community recommendation, which could evaluate the most present customers’information demand so as to reduce searching time and frequency during the information searching process. According to some authorized analysis, webpage information searching engine can be divided into two categories according to webpage content:directory webpage and theme webpage. The former is mainly constituted of links, which provides directory navigations; and the latter one mainly transmits concrete matter for customer in that the most content of it is pure text. After a deep analysis to customers before invitation reports and the basis of customer’s invitation pathway, this item could set up customer’s behavior model. After enough consideration of directory webpage’s characteristic, Q-greedy calculation and Markov link model is advocated so as these two one could reciprocal for each other; Theme webpage is mainly recommend resemble theme links for customer, actually which rest on information particle size theory, which could be divided into two particles so as to classify them deeply, finally it will recommend the high accurate information link for customer.According to the experiment’s results, a combination based on Q-greedy calculation and Markov link model based on customer’s invitation pathway, achieves a comparative satisfactory effect, moreover, recommendation of resemble theme webpage is also catering to customer’s demand and hoisted customers’satisfactory extent for searching information.