IEEE Transactions on Neural Networks
Equilibrium-Based Support Vector Machine for Semisupervised Classification
IEEE Transactions on Neural Networks
The geometrical learning of binary neural networks
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
In Organization the data is very important that increase the volume of information that is available on the web and that leads to the design of efficient and accurate web data classification systems. In this paper, we define a framework to improve the performance of a base classifier, by clustering the unlabeled data with labeled data using clustering algorithm (training of samples) labeling of clusters (majority voting for each Hyperspheres) and final generated classified data. We have used construction of BNN based semi-supervised classifier while training and testing of the classifier is performed. We have studied and customized a supervised classification algorithm to form out semi-supervised classification that leads to design a multiclass semi-supervised classifier using geometrical expansion. The experimental result shows provision for the classifier designer followed by training and testing medical disease dataset using pre-decided samples. Our classification model consists of training phase that covers two process clustering and labeling to perform classification task of medical data and the binary neural network is trained. In this we used two techniques normalization and quantization for pre-processing the datasets. Pre-processing impart various outcomes after applying the classification model like number of hypersphere, confusing samples that cannot be learned, training time and label of hypersphere. Comparison has been done for implementation and design of Binary Neural Network Classifier Algorithm with the other existing traditional algorithms. Our classifier evaluates performance in terms of generalization, number of hidden neuron and accuracy etc. The BNN-CA construct three-layered binary neural network (BNN) and can solve any semi-labeled multi-class problem.