为了提高自然语言网络答疑系统(Natural Language Web Answer System,简称NLWAS)中问答对的匹配速度和精度,提出了基于倒排词的倒排索引设计方案,详细阐述了索引体系结构、数据结构、匹配策略、动态维护,并对时间复杂度进行了分析。实际应用结果表明,使用倒排索引后系统性能优化效果明显增强。
In order to improve the matching speed and accuracy of the answer question pairs in Natural Language Web Answer System(NLWAS), propose an inverted index schema based on inverted word, introduce architecture of inverted index,data structure, matching strategy,and dynamic maintenance,also make an analysis of the time complexity in this paper. The experimental result shows that it has made prominent optimization of system design after the use of inverted index.