摘要
为了提高博客系统推荐标签的质量,分析了现有的标签推荐算法及相关技术,提出了一种基于社会化标注的博客标签推荐方法。该方法的优势在于:利用相似博客的社会化标签作为候选标签集,确保了推荐标签的全面性和可用性;基于TF-IDF相似度方法定义筛选步骤去除候选标签集中冗余和冷僻的标签,提高了推荐标签的准确性和高效性。实验结果表明了该方法的有效性。
To improve the quality of recommended tags in the bolgosphere, the algorithms of tag recommendation and related work are studied and a tag recommendation method for blogs based on social tagging is proposed. It has two advantages. First, to ensure the comprehensiveness and usability of the recommended tags, candidate tagset are selected from the tags of similar blogs based on the relationship between social tags and blog posts. Second, redundant tags and unfamiliar tags are got rid of from the candidate tagset based on TF-IDF similarity to improve the accuracy and high efficiency of the recommended tags. Finally, the experiments demonstrate the effectiveness of the method.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第12期4609-4613,共5页
Computer Engineering and Design
关键词
社会化标注
标签推荐算法
典型相关分析
文本特征加权方法
标签冗余
social tagging
tag recommendation
canonical correlation analysis
term frequency-inverse document frequency method
tag redundancy
