Authors : Niam Abdulmunim Al-Thanoon, Omar Saber Qasim and Zakariya Yahya Algamal
Abstract: Descriptor selection for classification methods is one of the most important topics in the chemometrics. The selection of descriptors can be considered to be a variable selection problem that aims to find a small subset of descriptors that has the most discriminative information for the classification target. Penalized Support Vector Machine (PSVM) is one of the most effective embedded methods and it is more preferable than the Support Vector Machine (SVM) because PSVM combines the standard SVM with a penalty to simultaneously perform both variable selection and classification. The PSVM with L1-norm is the most widely used methods. However, the efficiency of PSVM with L1-norm depends on appropriately choosing the tuning parameter which is involved in the L1-norm penalty. In this study, a particle swarm optimization method which is a metaheuristic continuous algorithm is proposed to determine the tuning parameter in PSVM with L1-norm penalty. The proposed method will efficiently help to find the most significant descriptors in constructing Quantitative Structure�Activity Relationship classification (QSAR) model with high classification performance. Depend on the four datasets, the experimental results show the favorable performance of the proposed method when the number of descriptors is high and the sample size is low comparing with other competitor methods.
Niam Abdulmunim Al-Thanoon, Omar Saber Qasim and Zakariya Yahya Algamal, 2020. Selection of Tuning Parameter in L1-Support Vector Machine via. Particle Swarm Optimization Method. Journal of Engineering and Applied Sciences, 15: 310-318.