论文部分内容阅读
【目的】研究专利检索日志中的同义词获取方法。【方法】提出一种基于用户行为分析的语义关系获取算法,利用检索式的逻辑运算符关系提取候选同义词对,结合拼音、字型、缩写、简繁等特征,从专利检索日志中挖掘出一部同义词词典。【结果】实验结果表明,该方法识别同义词的准确率达到74.5%,共生成17 495组同义词,生成词典的规模超过目前已有研究中的一些方法。【局限】该词典生成算法较适用于使用复杂检索式的图书情报检索领域。【结论】丰富了基于日志的语义词典获取领域的研究。
【Objective】 To study the method of obtaining synonyms in patent search journal. 【Method】 A semantic relationship acquisition algorithm based on user behavior analysis was proposed. Candidate synonym pairs were extracted from the relationship of logical operators of search queries. According to the characteristics of pinyin, fonts, abbreviations, Department Synonym Dictionary. 【Result】 The experimental results show that the accuracy of this method in identifying synonymous words reaches 74.5%, and 17 495 sets of synonymous words are generated. The size of the generated dictionaries exceeds that of the existing research. [Limitations] The dictionary generation algorithm is more suitable for the use of complex search library information retrieval field. 【Conclusion】 The study enriches the field of semantic dictionary acquisition based on log.