A Classification Method of Imbalanced Data Base on PSO Algorithm

• 作者: Chunkai Zhang Fengxing Shi Junru Lu
• 来源期刊: 国际计算机前沿大会会议论文集
• 学科: 社会科学
• 关键词: Imbalaced Re-sampling Sampling RATE PSO
• 年，卷（期）: 2016,（2）
• 页码范围: 37-39
• 页数: 3页
• 分类号: C5

摘要

In order to improve the performance of the resampling algorithm in imbalanced data learning and the optimal combination of the weight coefficients of the base classifiers in the ensemble classification algorithm,a new algorithm based on particle swarm optimization(PSO)is proposed.In traditional sampling method,sampling rate is often set artificially.This would make the final classification can’t get a optimal solution.By using the particle swarm optimization algorithm,the sampling rate of the boundary samples and the safety samples are optimized to obtain the optimal over sampling rate self-adaptivly,the weekness of setting sampling rate of traditional method would be overcome.