基于混合采样的非平衡数据集分类研究-计算机软件与理论专业论文.docxVIP

重庆大学 重庆大学硕士学位论文 英文摘要 PAGE PAGE IV concentrates the sampling area to samples which is difficult for distinguish, and increases the sampling area to samples which is difficult for distinguish, and increases the sampling rate for samples identified as security to make the classifier pay more attention to these samples hard to classify. 4. Improved the traditional random under-sampling strategy, and a boundary region cutting algorithm is proposed. By analyzing density and density reachable of negative sample dangerous points, applying under-sampling processing to the negative class samples, avoiding cleaning mistakenly for some important classes of negative samples in traditional algorithm. Keywords: hybrid-sampling, misclassified samples, unbalanced data, support vector machine 重庆大学硕 重庆大学硕士学位论文 目 录 目 录 中文摘要········································································································ I 英文摘要······································································································· II HYPERLINK \l _bookmark0 1 绪论 1 HYPERLINK \l _bookmark1 1.1 课题研究背景及意义 ····················································································1 HYPERLINK \l _bookmark2 1.2 本文主要研究内容及创新点 ···········································································2 HYPERLINK \l _bookmark3 1.3 本文的组织结构 ··························································································2 HYPERLINK \l _bookmark4 2 非平衡数据集分类的研究综述 ·······························································4 HYPERLINK \l _bookmark5 2.1 非平衡数据集分类的难点 ··············································································4 HYPERLINK \l _bookmark6 2.2 非平衡数据集分类的相关研究 ········································································6 HYPERLINK \l _bookmark7 2.2.1 算法层面的方法 ··································································