| 基于改进VSM的中文问答系统研究 |
| |
| 引用本文： | 苏小虎,杨思春.基于改进VSM的中文问答系统研究[J].情报理论与实践,2008,31(4). |
| |
| 作者姓名： | 苏小虎 杨思春 |
| |
| 作者单位： | 安徽工业大学,计算机学院,安徽,马鞍山,243002 |
| |
| 基金项目： | 安徽省高校省级自然科学基金项目研究成果之一,项目 |
| |
| 摘 要： | 针对向量空间模型中的权重计算公式仅考虑词语项在文档中的相关频数,提出词语项本身的领域权重概念,改进了向量空间模型的权重计算.同时结合关键词距离和关键词顺序信息,实现了句子相似度计算,以特定课程的FAQ库检索作S@n测试对比,结果表明改进后的相似度模型提高了s@n值
|
| 关 键 词： | 向量空间模型 领域权重 中文问答系统 |
Research on the Improved VSM-based Chinese Question Answering System |
| |
| Su Xiaohu et al.Research on the Improved VSM-based Chinese Question Answering System[J].Information Studies:Theory & Application,2008,31(4). |
| |
| Authors: | Su Xiaohu |
| |
| Abstract: | In view of the fact that only the term's relevant frequency in documents is considered in the weight calculation formula of Vector Space Model(VSM),a concept of term's domain weight is put forward to improve the weight calculation of VSM.Further more,the Keywords'distance and sequence are combined to realize the similarity calculation of the sentence.By conducting the S@n test with the special course's FAQ database,the results show that the S@n value is increased by the improved similarity model. |
| |
| Keywords: | VSM domain weight Chinese Question Answering System |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
