基于K-means的机构归一化研究 |
投稿时间:2013-01-08 点此下载全文 |
引用本文:孙海霞,李军莲,吴英杰.基于K-means的机构归一化研究[J].医学信息学杂志,2013,34(7):41-44 |
摘要点击次数: |
全文下载次数: |
|
基金项目:国家“十二五”科技支撑计划项目“科技知识组织体系的协同工作系统和辅助工具开发”(项目编号:2011BAH10B02) |
|
中文摘要:分析k-means算法的核心思想和基本步骤,借鉴现有基于频繁词集的文本聚类初始中心确定方法,提出一种面向大规模机构名称归一化处理应用的机构聚类方法,详细阐述机构聚类中心的生成、相似度算法的选择以及迭代次数问题,其实验和应用效果表现良好。 |
中文关键词:机构归一 机构聚类 K-means 频繁词集 相似度计算 |
|
Research on Institutions Normalization Based on K-means |
|
|
Abstract:The paper analyzes the core idea and basic steps of k-means, learns from the existing methods of determining initial text cluster centers which is based on frequent word sets, proposes a practical institutions cluster method meeting the need of large-scale institutions processing applications. It concretely elaborates the generation of institution clustering center, the selection of similarity algorithm and iterative times, the experimental results and its application perform well. |
keywords:Institutions normalization Institutions cluster K-means Frequent word sets Similarity computation |
查看全文 查看/发表评论 下载PDF阅读器 |