A CUDA Implementation of Random Forests: Early Results
Blekinge Institute of Technology, School of Computing2010 (English)Conference paper (Refereed) Published
Machine learning algorithms are frequently applied in data mining applications. Many of the tasks in this domain concern high-dimensional data. Consequently, these tasks are often complex and computationally expensive. This paper presents a GPU-based parallel implementation of the Random Forests algorithm. In contrast to previous work, the proposed algorithm is based on the compute unified device architecture (CUDA). An experimental comparison between the CUDA-based algorithm (CudaRF), and state-of-the-art parallel (FastRF) and sequential (LibRF) Random forests algorithms shows that CudaRF outperforms both FastRF and LibRF for the studied classification task.
Place, publisher, year, edition, pages
Göteborg: Chalmers Institute of Technology , 2010.
machine learning, graphics processing unit, random forests
IdentifiersURN: urn:nbn:se:bth-7705Local ID: oai:bth.se:forskinfo7C7A825038DE6570C12577E3004FFB7COAI: oai:DiVA.org:bth-7705DiVA: diva2:835353
Third Swedish Workshop on Multi-core Computing