【摘要】 Tag是一种对网络日志进行个性化整理的民间分类方式,具有平面化、社会化、人性化以及随意性、聚合性、自适应性等特点。使用Tag既能记录用户的关注点,还能提高检索的准确性等。本文在构建基于Tag的个性化信息检索系统时,一是通过引入聚类分析、归并等后控手段,提高了系统检索的准确度和匹配度;二是通过给系统模块添加智能因素,推测用户兴趣偏好,实现了系统与用户的交互和学习。尤为重要的是,笔者利用加权平均等做法较好地克服了Tag作为原生态自然语言存在的诸如模糊性、相似性、无意义性等缺点,使本系统更加个性化、智能化、实用化。更多还原

【Abstract】 Tag is a civil classification mode used for personized arrangment of network log,It has characteristics of planarizartion,socilization,personalization as well as causulness,collectivity and adaptability.Tag mode can both record the core concern of users and inprove retrieval accurary.In this article,when establishing the personalized information retrieval system based on Tag,the new mode for one thing enhences accurary and match degree by introducing post-controlled methods,such as clustering analysis and merging;on another it makes conjecture on the user’s interest and preference by means of adding intelligent factors,and realizes the interaction and study between system and user.What appears particularly important is that the author uses the weighted averages method to overcome many defects of Tag exsited as an original natrual language,such as fuzzness,similarity,and pointlessness,and manages to make the system more personalized,intellectualized and practical.更多还原