A GPU Based SVM Method with Accelerated Kernel Matrix Calculation

Support vector machine (SVM) is a popular classifier dealing with small-scale datasets. It has outstanding performance compared to other classifiers. However the execution time is extremely long when training Big Data. The Graphics Processing Unit (GPU) is a massively parallel device which performs very well as a co-processor. NVIDIA proposed a programming platform, CUDA, in 2006, which makes it much easier to program on GPU for large-scale calculation. This paper proposes a GPU based accelerated SVM method. Matrix multiplication, which is the main procedure of kernel matrix calculation in the training stage, is the major issue for high time complexity. Thus, we propose a method to calculate the kernel matrix using GPU. The polynomial kernel function is applied in this paper. Experiment results indicate that the proposed method can increase the computation speed by as much as 41 times for the studied datasets.

[1]  Maozhen Li,et al.  An ontology enhanced parallel SVM for scalable spam filter training , 2013, Neurocomputing.

[2]  Daphne Koller,et al.  Support Vector Machine Active Learning with Applications to Text Classification , 2000, J. Mach. Learn. Res..

[3]  François Poulet,et al.  GPU-based Parallel SVM Algorithm , 2009 .

[4]  Ji-Bo Wang,et al.  GPU Accelerated Support Vector Machines for Mining High-Throughput Screening Data , 2009, J. Chem. Inf. Model..

[5]  B. Venkataramani,et al.  Study and evaluation of a multi-class SVM classifier using diminishing learning technique , 2010, Neurocomputing.

[6]  Sergio Herrero-Lopez,et al.  Accelerating SVMs by integrating GPUs into MapReduce clusters , 2011, 2011 IEEE International Conference on Systems, Man, and Cybernetics.

[7]  Koby Crammer,et al.  On the Algorithmic Implementation of Multiclass Kernel-based Vector Machines , 2002, J. Mach. Learn. Res..

[8]  Ioannis Kompatsiaris,et al.  GPU acceleration for support vector machines , 2011, WIAMIS 2011.

[9]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[10]  Dinesh Manocha,et al.  Query co-processing on commodity processors , 2006, VLDB.

[11]  Bo Yan,et al.  DVT-PKM: An Improved GPU Based Parallel K-Means Algorithm , 2014, ICIC.

[12]  Qi Li,et al.  Parallel multitask cross validation for Support Vector Machine using GPU , 2013, J. Parallel Distributed Comput..

[13]  Miao Xin,et al.  An Implementation of GPU Accelerated MapReduce: Using Hadoop with OpenCL for Data- and Compute-Intensive Jobs , 2012, 2012 International Joint Conference on Service Sciences.

[14]  Ivor W. Tsang,et al.  Core Vector Machines: Fast SVM Training on Very Large Data Sets , 2005, J. Mach. Learn. Res..

[15]  Jesse Patrick Harvey,et al.  GPU acceleration of object classification algorithms using NVIDIA CUDA , 2009 .

[16]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[17]  Jason D. M. Rennie,et al.  Improving Multiclass Text Classification with the Support Vector Machine , 2001 .