计算机与现代化

• 人工智能 • 上一篇    下一篇

基于跳跃显露模式挖掘算法的癌症分类

  

  1.   (河海大学计算机与信息学院,江苏南京211100)
  • 收稿日期:2018-03-09 出版日期:2018-06-13 发布日期:2018-06-13
  • 作者简介:乔媛(1992-),女,江苏南京人,河南大学计算机与信息学院硕士研究生,研究方向:信息处理与信息系统; 廖小平(1965-),男,江苏南京人,副教授,硕士,研究方向:信息处理与信息系统; 邵开霞(1992-),女,江苏南京人,硕士研究生,研究方向:大数据,数据管理。

 Cancer Classification Based on Jumping Emerging Pattern Mining Algorithm

  1.  (College of Computer and Information, Hohai University, Nanjing 211100, China)
  • Received:2018-03-09 Online:2018-06-13 Published:2018-06-13

摘要: 分类问题是数据挖掘中的一项重要课题,然而目前对于癌症数据的分类研究还相对较少。近年来提出的强跳跃显露模式SJEP是一种具有很强区分能力的新模式,对于癌症数据的分类具有明显的优势。为了使癌症数据的分类精确度得以进一步提升,本文引入集成学习的思想,对原有的Boosting算法做出一些改进,并将改进后的Boosting算法与SP-树分类算法相结合,提出一种以SP-树分类算法作为基学习算法的SP_Boost算法。

关键词: 分类算法, 强跳跃显露模式, 集成学习

Abstract: Classification is an important topic in data mining, but currently, the classification of cancer data is still relatively small. In recent years, Strong Jumping Emerging Patterns (SJEP) is a new model with strong ability to distinguish, which has obvious advantages for the classification of cancer data. In order to further improve the classification accuracy of cancer data, this paper introduces the idea of integrated learning, and makes some improvements to the original Boosting algorithm. The improved Boosting algorithm is combined with SP-tree classification algorithm, an SP_Boost algorithm based on SP-tree classification algorithm is proposed.

Key words: classification algorithms, strong jumping emerging patterns, ensemble learning

中图分类号: