Usability of KSOM and classical set in information retrieval

The Kohonen self-organizing feature map (KSOFM) also known the Kohonen self-organizing map (KSOM) under comes the artificial neural network. An important characteristic of ANN is ability to learn. Learning is a process by adjusting the weight with respect to input features so that the actual output response will matches to the desired output. The role of learning based network is that the point distributes themselves across the input space to selected or mapped group of similar input vectors, while the output nodes compete among themselves. Once network trained, the network produces a low dimensional data clustering of the input value that preserves the ordering of the actual structure. Thus nodes can be recognizing groups of similar input vectors. It generates a mapping of the input vector to the output layer, which depends on input vectors and overall result in dimensionality reduction of the input space. Follows the “winner-take-all” policy means which cluster unit weight matches more closely with the input pattern is considered the winning neurons so that pattern classification process take place. It is basically based on sumup of actual sequence and final junction point.