期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2012
卷号：3
期号：5
出版社：Seventh Sense Research Group
摘要：The main ability of neural network is to learn from its environment and to improve its performance through learning. For this purpose there are two types of learning supervised or active learning – learning with an external ‘teacher’ or a supervisor who present a training set to the network. But another type of learning also exists unsupervised learning[1] . Unsupervised learning is self organized learning doesn’t require an external teacher. During training session neural network receives a number of input patterns , discovers significant features in these patterns and learns how to classify input data into appropriate categories. It follows the neuro biological organization of the brain. These algorithms aim to learn rapidly so learn much faster than backpropagation networks and thus can be used in real time. Unsupervised NN are effective in dealing with unexpected and changing conditions[3]. There are basically two major self – organising networks based learning Hebbian and competitive learning. We will use Hebbian learning in this paper to visualize that how it can help in feature extraction from any data. We will use input vector weight matrix examples and denote presence of a feature by 1 and absence of a feature by 0. In this method we will see that how features are identified and we can discover patterns in given data. We can use this method for classification and clustering purpose also. Then we will apply this learning rule on web data or content for discovers pattern & extraction of features. Weight increases when same pattern repeats and decrease when it doesn’t repeat. The network associates some input xi with some outputs yi and yj because input xi and xj coupled during training. But it cannot associate some input x with some output y because that input didn’t appear during training and our network has lost the ability to recognize it.
