TY - GEN
T1 - A new SVM model for classifying genetic data
AU - Chen, Wang Juh
AU - Guo, Hongbin
AU - Renaut, Rosemary
AU - Chen, Kewei
PY - 2010
Y1 - 2010
N2 - We propose a new formulation of the Support VectorMachine (SVM) for classifying genetic data. It is based on the development of ideas from the method of total least squares, in which assumed error in measured data are incorporated in the model design. For genetic data the number of features is always far greater than the sample size. Consequently, in our method, we introduce Lagrange multipliers and solve for the dual variables. Instead of finding the minimum value of the Lagrangian function, we solve the nonlinear system of equations obtained from the Karush-Kuhn-Tucker conditions. We also implement complementarity constraints and incorporate weighting of the linear system by the inverse covariance matrix of the measured data. The proposed algorithm gives improved results and higher sensitivity for classifying a set of Alzheimer's Disease Positron Emission Tomography images as compared with SVM. It is also more robust to noise than SVM. Copyright
AB - We propose a new formulation of the Support VectorMachine (SVM) for classifying genetic data. It is based on the development of ideas from the method of total least squares, in which assumed error in measured data are incorporated in the model design. For genetic data the number of features is always far greater than the sample size. Consequently, in our method, we introduce Lagrange multipliers and solve for the dual variables. Instead of finding the minimum value of the Lagrangian function, we solve the nonlinear system of equations obtained from the Karush-Kuhn-Tucker conditions. We also implement complementarity constraints and incorporate weighting of the linear system by the inverse covariance matrix of the measured data. The proposed algorithm gives improved results and higher sensitivity for classifying a set of Alzheimer's Disease Positron Emission Tomography images as compared with SVM. It is also more robust to noise than SVM. Copyright
UR - http://www.scopus.com/inward/record.url?scp=84878121643&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84878121643&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84878121643
SN - 9781617820694
T3 - International Conference on Bioinformatics, Computational Biology, Genomics and Chemoinformatics 2010, BCBGC 2010
SP - 54
EP - 60
BT - International Conference on Bioinformatics, Computational Biology, Genomics and Chemoinformatics 2010, BCBGC 2010
T2 - 2010 International Conference on Bioinformatics, Computational Biology, Genomics and Chemoinformatics, BCBGC 2010
Y2 - 12 July 2010 through 14 July 2010
ER -