A Meta-Heuristic Model for Data Classification Using Target Optimization

2. Related Work

The recent studies have been stated that, a supervised fuzzy logic based adaptive resonance theory neural network called as FAM, which was incorporated with GSA for the classification problem (Tan et al., 2015). They have evolved the connection weights of the FAM with GSA. It has also used a hybrid approach for classification problem which is combined chaotic search with GSA and has used them in the training process of SVM. GSA has been used to determine the real valued parameters of SVM (Li et al., 2015). A Hybrid Neural Network and Gravitational Search Algorithm (HNGSA): a fusion model based on neural network and GSA, has been proposed for a solution of Wessinger’s equation (Kumar, 2014; Ghalambaz et al., 2011). Here, the role of GSA is for providing training to a multi-layer feed-forward neural network. It has integrated support vector machine (SVM) with GSA to increase the classification accuracy in binary class problem and to avoid the parameter fixing problem in SVM. They had taken the performance matrices as Standard deviation, Mean and Max.