Scholarlink

This study establishes a comprehensive classification system for user search assistance needs through extensive research on existing studies and in-depth interviews with search users. Subsequently, based on the proposed auxiliary demand classification, three machine learning algorithms and a deep learning algorithm are applied to construct a classification prediction model, which implements a classification prediction method from user search behavior characteristics to user search auxiliary demand. The experimental results show that the random forest model exhibits the best performance in the prediction problem, and its average prediction accuracy and AUC index for the six categories of search assistance auxiliary needs are close to 0.9.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

User Search Assistance in Information Retrieval: Building a Classification System and Predicting User Needs

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yiming Zhao,
  • Zhan Chen,
  • Fan Zhang,
  • Jiani Wu,
  • Yanming Chen,
  • Qiuqi Xie,
  • Shuaiju Yu

摘要

This study establishes a comprehensive classification system for user search assistance needs through extensive research on existing studies and in-depth interviews with search users. Subsequently, based on the proposed auxiliary demand classification, three machine learning algorithms and a deep learning algorithm are applied to construct a classification prediction model, which implements a classification prediction method from user search behavior characteristics to user search auxiliary demand. The experimental results show that the random forest model exhibits the best performance in the prediction problem, and its average prediction accuracy and AUC index for the six categories of search assistance auxiliary needs are close to 0.9.