Research On Adaptive Blog Recommendation Based On Content

Ever since the Internet began to change, the blog also slowly rises at an alarming rate.So far, the blog users have reached millions. The blog has become an indispensable toolfor people’s working and living, and the rapid increase in the number of users make theblog posting increase naturally. As blog resources are gradually richer, more and moreusers want to solve difficult problems and look up relevant information through blogs.But the results recommended are simply chosen by clicking rate and reading amount,and it is not according to the user actual needs to provide users with the relevant blog.Therefore, in order to improve blog-adhesive when the user using the blog, user needsneed to be analyzed and user’s interests have to be known and users’ interests areaccurately grasped. When the user’s interests tend changes, recommendation needs to beadjusted actively in order to meet the user’s needs.According to the blog users changeability of interest, a method is proposed based ona user feedback adaptive recommend-heuristic. Firstly, through the analysis of the blogtype that users post to establish user model, based on this the reading behavior of otherusers is recorded; And then analysis of user feedback records, we mine the state of mindfrom the user actions, and record user reading the blog analysis way from the user’snavigation records, also track the user’s interest points, while testing whether the resultswill recommend the needs of users, to adjust the recommended results; When interest inthe user model change, we use a bayes correction method to update user model.The experiment is divided into the method comparison experiment with the thresholdsettings and recommendation base choice experiment. The recommended base of Thethreshold set method is5-50. As the increase of the recommended base, the number ofblog recommended increases, and the user’s interest will be more comprehensive, whichalso should be recommended more comprehensive and more accurate. But it willinfluence the number of blogs user’s reading habits, so the user will not pay muchattention to scroll the recommended after content. In the early stage of the user readinghave ascension, along with excessive number will gradually decrease recommend effect.Recommend base is through the different choice experiment settings to look for a basebest recommend base, and in different environment is not recommended base.Experiments show that the heuristic recommend method can effectively increase theadaptive recommend the quality of the blog.