基于排序集成的自动术语识别方法

摘要

Automatic term recognition ( ATR) is one of the key steps in the fields of information extraction and text mining. As the basic ATR algorithms use characteristic information of certain aspects, they have noticeable limitations. The paper introduces the local Kemeny optimal method to deal with ATR issue, and presents a new rank aggregation method. Experiment results show that the method has significantly improved the accuracy of automatic term recognition.%自动术语识别是信息抽取和文本挖掘等领域的关键步骤之一.基础自动术语识别算法采用某些方面的特征信息,有明显的局限性,引入局部Kemeny最优的方法来处理自动术语识别问题,并提出新的集成方法.实验结果表明该方法显著改善了自动术语识别的精准度.