基于网页主题相关度和标签相似度的改进PageRank算法研究

Research on Improved PageRank Algorithm Based on Web Subject Relevance and Tag Similarity

文章将PageRank算法与社会化标签进行结合,提出一种基于链接网页主题之间相关度和社会化标签之间相似度的改进PageRank算法.首先基于信息特征词构建向量空间模型,通过余弦值和TF-IDF算法计算网页主题相关度;然后建立社会化标签向量计算链接网页标签相似度;最后确定权重关系进行算法迭代,从而实现Web页面的重新排序.实验表明,该算法能提高信息推荐的准确性,但算法质量不稳定,推荐效果呈下降趋势.

Combining PageRank algorithm and social tags,this paper proposes an improved PageRank algorithm based on the relevance between each linked web page topic and the similarity between social tags.A vector space model based on information feature words is established,and the topic relevance of web pages is calculated by cosine and TF-IDF algorithm.Social tags vector are then established to calculate the relevance of linked web page tags,and finally the weight is determined to iterate the algorithm to reorder the Web pages.Experimental results show that this algorithm can improve the accuracy of information recommendation,but the quality of the algorithm is unstable so the recommendation effect is decreasing.