基于隐马尔科夫模型的浏览路径预测

Research prediction based on hidden Markov model

基于马尔科夫模型的浏览路径预测,仅仅从用户的浏览会话本身出发来预测用户下一步的链接,并不能捕获用户的真正兴趣所在。运用隐马尔科夫模型来分析用户浏览网页的内容,可进一步捕获用户的浏览兴趣,并作下一步的链接预测。当浏览序列长度逐渐增加,系统捕获的用户浏览信息越来越多,此时能够折射出用户的兴趣所在,预测准确率也逐步增加。当浏览序列长度大于或等于8时,预测准确率已经到达80%,提高了预测准确率。

Research prediction of based on Latent Markov model, only from to research conversation itself set out to predict user's next chaining user, can not catch users' real interest. This text use Hidden Markov Model, analyse users browse through the content of the webpage, catch users' interest of research further, do next chaining to predict, When the browsing sequence length increases gradually, the system capture user browsing information are more and more many, this time can refract the user the interest to be at, forecast the rate of accuracy also gradually increases. When the browsing sequence length is bigger than or was equal to 8, forecast the rate of accuracy already arrived 80%, have improved the rate of accuracy of predicting.