Multi Objective Artificial Immune System Weighted Feature Selection for Classification

Supervised learning algorithm based on an AIS using clonal selection is already proposed [1]. Even though some AIS for classification have been developed, an artificial immune system for classification with local feature selection model has a unique feature; it includes the local feature selection mechanism. The aim of feature selection is to reduce the dimension of the input vector by the selection of a feature (variable) subset which describes the object in the best manner and ensures the best quality of the learning model. This set can be reduced during an optional apoptosis process. Local feature selection and apoptosis result in data-reduction capabilities. The classifier has only two user-settable parameters controlling the global-local properties of the feature space searching. In order to increase the performance of the system even more, in this paper proposed the weighted feature selection technique. In the weighted feature selection approach, we assign the weight value for each and every features based on their importance. Due to this technique, we can further reduce the features of the classification process. It is used to determine the levels of importance of the features (the strength of individual feature binding). For assigning weights to features, in other words a feature with weights proportional to the identification accuracy of individual features. Here assign the proper weight to each feature is helpful to improve the accuracy of the classification. The proposed system is effective in the feature reduction and has high classification accuracy compared to the existing system.