摘要
以名人网页语料库为基础,研究名人网页的个性化检索设计,实现了名人网页相关度评价的概率模型.从模型的训练策略、相关度计算公式等方面进行实验分析,总结了概率模型的规律,提出了模型的改进方法.实验表明,通过选择合适的训练集、细化用户反馈信息,模型取得了较高的相关度评价准确率.
Based on the corpus of Chinese celebrities,Designs and implements the probabilistic model for relevance evaluation of celebrities' webpages, the main process of personalized information retrieval of celebrities' webpages. After analyzing on the experiment results from aspects such as the training strategy of the model, relevance computing formula and so on, Achieves some valuable conclusions about the probabilistic model and how to optimize the model. As is shown in the experiment, with well selected training sets and more detailed user feedback, the evaluation precision reaches a high level.
出处
《河南教育学院学报(自然科学版)》
2008年第1期39-41,共3页
Journal of Henan Institute of Education(Natural Science Edition)
关键词
个性化检索
概率模型
相关度评价
personalized information retrieval
probabilistic model
relevance evaluation