An improved weighted extreme learning machine for imbalanced data classification
| Autor: | Gao-Yan Zhang, Ying Mei, Chengbo Lu, Huihui Xu, Hai-Feng Ke |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
0209 industrial biotechnology
Majority rule Control and Optimization General Computer Science Computer science media_common.quotation_subject 02 engineering and technology Machine learning computer.software_genre Imbalanced data 020901 industrial engineering & automation Voting 0202 electrical engineering electronic engineering information engineering media_common Extreme learning machine Randomized weighted majority algorithm Weighted Majority Algorithm business.industry Pattern recognition Statistical classification ComputingMethodologies_PATTERNRECOGNITION Classification result 020201 artificial intelligence & image processing Artificial intelligence business computer |
| Zdroj: | Memetic Computing. 11:27-34 |
| ISSN: | 1865-9292 1865-9284 |
| Popis: | This paper proposes an improved weighted extreme learning machine (IW-ELM) for imbalanced data classification. By incorporating voting method into weighted extreme learning machine (weighted ELM), three major steps are involved in the proposed method: training weighted ELM classifiers, eliminating unusable classifies to determine proper classifiers for voting, and finally determining the classification result based on majority voting. Simulations on many real world imbalanced datasets with various imbalance ratios have demonstrated that the proposed algorithm outperforms weighted ELM and other related classification algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink