Challenges in Evolutionary Algorithm to Find Optimal Parameters of SVM

In rapidly changing classification and predation environment, optimization techniques in determining the hyper parameter of Support Vector Machine has become crucial for the accuracy of result. It’s an important tool for improving output quality of classification and prediction which includes modeling parameters relationship and resolution of optimal hyper parameter. However, determination of regularization (C) and gamma (γ), through mathematical models have undergone substantial development and expansion. In this paper, optimization techniques are categorized under several criteria. We also included the benchmarks for measuring the performance of classifier after parameter tuning, and found there is still a scope of improvement

Algorithm · Evolutionary algorithm · Machine learning · Mathematical optimization · Support vector machine · Computer Science · Evolutionary Algorithms and Applications · Imbalanced Data Classification Techniques · Mathematics · Metaheuristic Optimization Algorithms Research · Artificial Intelligence

• An introduction to ROC analysis

  Open Access•Tom Fawcett•Pattern Recognition Letters•2006

  Incluir item

• No free lunch theorems for optimization

  Open Access•David H Wolpert, William G Macready•IEEE Transactions on Evolutionary…•1997

  Incluir item

• Libsvm

  Open Access•Chih-Chung Chang, Chih-Jen Lin et al.•ACM Transactions on Intelligent…•2011

  Incluir item

• Grey Wolf Optimizer

  Seyedali Mirjalili, Seyed Mohammad Mirjalili et al.•Advances in Engineering Software•2014

  Incluir item

• Support-Vector Networks

  Open Access•Corinna Cortes, Vladimir Vapnik•Machine Learning•1995

  Incluir item

• The meaning and use of the area under a receiver operating characteristic (ROC) curve.

  J A Hanley, Barbara J Mcneil•Radiology•1982

  Incluir item