Scholarlink

<p>This study presents a new data stream learning framework for class imbalance data sources that can improve classification. This framework uses a genetic algorithm to select the best features and perform fast resampling using intelligent techniques to identify clusters with unique inheriting structures. The proposed approach learns one chunk at a time without requiring access to previous data and emphasizes updating misclassified examples in the model building. The experiments demonstrated that the suggested framework improves performance for classifying the imblance data stream. Moreover, the proposed approach exhibited the best performance on compared models with different evaluation metrics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel framework for improving class imbalance learning using feature space identification and fast informative resampling techniques

  • 期刊论文

  • 发表日期：

    2025年5月11日

查看全文
  • V. S. NageswaraRao Kadiyala,
  • Sabuzima Nayak,
  • Ripon Patgiri

摘要

This study presents a new data stream learning framework for class imbalance data sources that can improve classification. This framework uses a genetic algorithm to select the best features and perform fast resampling using intelligent techniques to identify clusters with unique inheriting structures. The proposed approach learns one chunk at a time without requiring access to previous data and emphasizes updating misclassified examples in the model building. The experiments demonstrated that the suggested framework improves performance for classifying the imblance data stream. Moreover, the proposed approach exhibited the best performance on compared models with different evaluation metrics.