摘要
[目的/意义]术语是本体的重要组成部分,术语自动抽取是本体自动构建的基础,文章采用回归的方法对未登录词进行概率(某个数值(组合)对应的候选词集合中术语的概率)预测,获得该词可能为术语的概率。[方法/过程]文章结合语言学和统计方法,通过构建术语库提取术语抽取模板来抽取候选术语,此外,通过引入回归的方法,将术语抽取问题转化为对词语成为术语的概率的预测问题。[结果/结论]提出的方法最后通过实验验证了其有效性。
[ Purpose/significance ] Terminology is one of the most important parts of ontology. Automatic terminology extraction is the basis of automatic ontology construction. This paper adopts the method of regression to predict the probability of unknown words （ a value or combined values corresponds to the set of candidate words in terms of probability） . [ Method/process] The pa- per focuses on the combination of linguistics and statistical methods to extract the candidates based on rule-based method. By using the method of regression, the paper takes the terminology extraction problem as the prediction issue of terminology probability. [ Result/conclusion ] The experiment verifies the validation of the proposed method.
出处
《情报理论与实践》
CSSCI
北大核心
2016年第1期24-31,15,共9页
Information Studies:Theory & Application
基金
国家社会科学基金重大招标项目"面向学科领域的网络信息资源深度聚合与服务研究"的成果
项目编号:12&ZD221
