论文部分内容阅读
[目的/意义]提出一种基于潜在语义分析和最大边缘相关性的方法,向用户推荐具有高相关多样性的学术论文。[方法/过程]首先采用潜在语义分析方法处理高维词项—文档空间和用户兴趣向量,取相似度大于阈值的文献形成备选相关文献集;然后根据信息新颖度的思想,利用引文分析方法计算备选文献的差异性;最后使用最大边缘相关性方法处理备选相关文献,形成冗余度小相关多样性高的推荐论文集。[结果/结论]实验结果表明,该方法的推荐准确率和多样性优于其他两种基准方法。[局限]随着文本数据维数增加,算法的时间复杂度增加,耗时增加。
[Purpose / Significance] A method based on latent semantic analysis and maximum edge correlation is proposed to recommend to users a highly relevant scholarly dissertation. [Methods / Processes] Firstly, the latent semantic analysis method is used to deal with the high-dimensional term items - document space and user interest vectors, and the documents with similarities greater than the threshold value are selected to form alternative related documents. Based on the novelty of information, citation analysis Calculate the difference of alternative documents; finally, use the method of maximum edge correlation to deal with alternative relevant documents, and form a set of recommended papers with high degree of redundancy and high correlation diversity. [Results / Conclusions] The experimental results show that the proposed method has better accuracy and diversity than the other two benchmark methods. [Limitations] As the number of text data increases, the time complexity of the algorithm increases and the time-consuming increases.