Scholarlink

An imbalanced dataset refers to a dataset where the number of samples in each category varies significantly. The training of unbalanced data will bring problems such as large classification error, so it should be balanced. ADASYN algorithm, AHC algorithm, SMOTE algorithm and ADOMS algorithm are the most commonly used methods for oversampling unbalanced data. After balancing the data, the existing noise problems will affect the effect and performance of data classification. This paper combines the automatic denoising algorithm of neural network with the above four algorithms, and verifies its effect through experiments. Experiments show that the combination of ADOMS + neural network automatic denoising algorithm can reduce the noise of balanced data set by 5.1%, and the effect is better.

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

Research on Imbalanced Datasets Classification Based on Neural Network and Oversampling Methods

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Peijin Zhu

摘要

An imbalanced dataset refers to a dataset where the number of samples in each category varies significantly. The training of unbalanced data will bring problems such as large classification error, so it should be balanced. ADASYN algorithm, AHC algorithm, SMOTE algorithm and ADOMS algorithm are the most commonly used methods for oversampling unbalanced data. After balancing the data, the existing noise problems will affect the effect and performance of data classification. This paper combines the automatic denoising algorithm of neural network with the above four algorithms, and verifies its effect through experiments. Experiments show that the combination of ADOMS + neural network automatic denoising algorithm can reduce the noise of balanced data set by 5.1%, and the effect is better.