Design a Binary Neural Network Classifier Algorithm with Parallel Training in Hidden Layer

-Since last decade, classification methods are useful in a wide range of applications. Classification is a task to group the sample having similar properties. This capability can be introduced in computer system by designing various types of classifiers. Neural network is one of the artificial intelligent techniques that have many successful examples when applying to this problem. Here, a Binary Neural Network Classifier (BNNC) is analyzed and implemented for solving multi class problem. It takes inputs in binary form and generates outputs in binary form. It forms three layered network architecture [1], first layer is an input layer, second layer is a hidden layer & last layer is an output layer. Firstly, it preprocesses data set for the sake of generating binary values. Then, preprocessed data is used for hidden layer as an inputs, the hidden layer training is done in parallel for all multiple classes to reduce the time for training. The BNNC offers high degree of parallelism in hidden layer formation.Each module in the hidden layer of BNNC is exposed to the patterns of only one class .The learning method is an iterative process to optimize the classifier parameters. In this approach, overlapping problem is tackled to enhance the performance of classifier. This is done by changing hypersphere radius. The output of hidden layer is combined at the output layer. This approach is tested with various benchmark datasets. The results corresponding to these datasets are generated by varying the learning parameters. It is found that accuracies for various dataset are varying from 60% to 93%. It is compared with constructive semi-supervised classifier Algorithm (CS-SCA) by varying different parameter values. Inmost of the cases accuracies improves in BNNC. This is due to the elimination of samples which are lying in overlapping region of classes in case of CS-SCA. Thus the overlapping issue here improves performance of this classifier. Keywords-SEMI SUPERVISED CLASSIFICATION, GEOMETRICAL EXPA-NSION, BINARY NEURAL NETWORK.