作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2007, Vol. 33 ›› Issue (04): 165-167. doi: 10.3969/j.issn.1000-3428.2007.04.057

• 人工智能及识别技术 • 上一篇    下一篇

CBR检索的线性回归模型及实现

刘缵敏1,2,孙 义1,史忠植2   

  1. (1. 北京科技大学计算机系,北京 100083;2. 中国科学院计算技术研究所,北京 100080)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-02-20 发布日期:2007-02-20

Linear-regression Model of CBR Retrieve and Implementation

LIU Zuanmin1,2, SUN Yi1, SHI Zhongzhi2   

  1. (1. Department of Computer Science, University of Science and Technology Beijing, Beijing 100083; 2. Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080)
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-02-20 Published:2007-02-20

摘要: 通过对CBR传统模型的分析与研究,针对传统CBR检索中主观确定特征权重的不足,提出了CBR检索的线性回归模型,该模型利用最小二乘法的线性回归性,更加科学、准确地确定各特征的权重,依据成熟的距离公式准确地求出范例的相似度,达到范例准确高效重用的目的。最后介绍了模型的实现方法,并且给出了详细的模型参数。

关键词: CBR, 检索, 回归, 最小二乘法, 权重

Abstract: After studying the traditional CBR model, according to the deficiency that the feature weights are subjectively defined, this paper proposes a linear-regression model for CBR retrieval, which will improve the effectiveness of CBR retrieval model. The key idea is to decide the weight of each feature by the method of least square. And its property for linear regression helps to make the weights more exact. Thus in the model the similarity degree between cases is more precise than the traditional one, which facilitates the reuse of the existing cases greatly. It also gives one method to implement the model and describes the parameters.

Key words: Case-based reasoning, Retrieval, Regression, Least square, Weight