基于概率潜在语义分析的Web用户会话识别

Web Session Identification Based on Probabilistic Latent Semantic Analysis

摘要: 为提高Web用户会话识别的准确性,给出了一种基于概率潜在语义分析模型(PLSA)和竞争奖罚(CRP)算法的Web用户会话识别方法.其核心内容是利用PLSA模型计算出请求页面和每一个活动用户会话的概率,比较概率判定请求页面应该归属的用户会话,并采用竞争奖罚算法判别用户会话的结束.实验结果表明:基于PLSA模型和竞争奖罚算法的用户会话识别方法的识别成功率高于其他常用的会话识别方法.

Abstract: To improve the validity and efficiency of the user session identification, a new approach of user session identification is presented, which is based on the probabilistic latent semantic analysis model and the competitive reward and punishment algorithm.The probabilistic latent semantic analysis model is used to compute the probabilities of a request and each active session.The model identifies which active session a request should belong to in terms of the probabilities.A competitive reward and punishment method is applied to determine the end of sessions.The experimental results show that the average percentage of successfully reconstructed sessions based on PLSA and CRP is higher than other session identification algorithms.