Using Kohonen networks for ship radio station identification

Neural networks are a tool with many different applications. They are used in approximation, control, compression, and identification tasks. Object identification can be performed by many different neural networks. One of them is the Kohonen network. It is a network that learns through self-organization of neurons, and its main application is data clustering and compression. Another application of the Kohonen network is identification.Using this network for object identification involves the need to use specialized learning algorithms. One such algorithm is the adaptive vector quantization algorithm. To test the effectiveness of the Kohonen network trained with the mentioned algorithm, a series of studies were conducted during which the network’s task was to identify ship radio stations. The results of the studies are presented at the end of the article. The classic Kohonen network is a self-organizing network most often associated with data grouping (clustering) tasks. (…)

Using Kohonen networks for ship radio station identification – PDF file for download