A Multi-Integrated Optimization Technique for Classification of Breast Cancer Disease Using an Artificial Neural Network
Abstract
Classification in data mining plays an important role in biomedical analysis that gives a better way to predict or diagnose any disease at an early stage. As the data size increases, the methodologies and techniques for selecting and optimizing the features also increases. It plays a challenging task for developing a computationally efficient and accurate predictive model. The proposed work presented the multi-integrated optimization method using Genetic algorithm (GA) and particle swarm optimization (PSO) and classifies using the artificial neural network (ANN). It presents an efficient approach for predicting cancer disease and also proved their efficiency and accuracy over other single optimization techniques. This system takes the union of both the optimal relevant features generating from PSO and GA.Thus, a new subset of data formed after two ways integrated feature selection reduces the complexity and enhances the performances when classifying with the ANN. The comparison of performance is also evaluated with different subsets of features after and before feature optimization. This experiment provides a better approach for data selection and gives an accuracy of 98.24% respectively.