摘要： 针对传统推荐系统中存在的冷启动、数据稀疏、语义缺乏、推荐精度较低等问题,提出一种基于事件本体的推荐算法。结合新闻的分类结构和新闻语料构建事件本体,对用户浏览的新闻进行要素抽取并构建用户兴趣模型。基于事件本体的分类结构计算新闻事件之间的相似度,通过用户兴趣模型计算用户兴趣相似度,根据事件本体非层次结构的语义半径寻找相关新闻事件。综合事件本体相似度、用户兴趣相似度和非层次结构相似度3个方面得出新闻个性化推荐结果。实验结果表明,该算法的推荐结果优于协同过滤推荐算法和基于内容的推荐算法。

关键词: 事件本体, 推荐系统, 新闻个性化推荐, 本体相似度, 用户兴趣相似度, 非层次结构相似度

Abstract: Aiming at the problems existing in the traditional recommendation system,such as cold start,sparse data,lack of semantics and relatively low recommendation accuracy,a recommendation algorithm based on event ontology is proposed.By combining the news classification structure and news corpus to build event ontology,the elements of news browsed by users are extracted and user’s interest model is constructed.The similarity between the news events is calculated based on the event ontology structure,the user interest similarity is calculated through user’s interest model,and relevant news events are found according to the semantic radius of the non-hierarchical structure of event ontology.Synthesize the news ontology similarity,the user’s interest similarity,and the non-hierarchical structure similarity to realize a comprehensive news personalized recommendation.Experimental results show that the proposed algorithm has better recommendation than the collaborative filtering recommendation algorithm and the recommendation algorithm based on content.

Key words: event ontology, recommendation system, news personalized recommendation, ontology similarity, user’s interest similarity, non-hierarchical structure similarity

中图分类号: 

• TP393