摘要
讨论了非结构化文档信息的检索模型,分析了传统交互式信息检索方法,提出了面向基金项目文档查询的智能交互式信息检索过程和处理流程。基于用户评价项目文档的反馈信息,采用ID3算法、CLCC算法和SVM分类函数分别学习用户查询的潜在意图和目标,并应用所学习的规则知识或分类函数支持项目文档查询。以某基金评审管理中项目文档的查询为例进行了实验计算与分析。
Firstly it analyses the importance of documents retrieval in the evaluation of fund projects,and discusses the conventional interactive information retrieval method for non-structured documents. Then it puts forward the intelligent and interactive retrieval procedure and function for searching fund project documents. The user’s evaluation information on retrieval documents is fed back,and is further used to learn the potential search intention and aim of the user through implementing machine learning methods as ID3,CLCC algorithm,and SVM classification function. The learned rule knowledge or classification functions are adopted to support the retrieval of project documents. The proposed method is applied to the retrieval of a specific fund project documents database experimentally.
出处
《研究与发展管理》
CSSCI
北大核心
2005年第3期106-110,共5页
R&D Management
基金
高等学校博士学科点专项科研基金资助(20020056047)。
关键词
基金评审管理
机器学习
项目文档检索
智能交互式检索
evaluation management of fund projects
machine learning
project document retrieval
intelligent and interactive information retrieval
