... as a three layer back propagation type; the meaning of this is explained later. As part of the investigation, the network was trained, using a random noise generating function to vary the pattern slightly each time. The patterns used were two dimensional grids of ones and zeros, depicting horizontal, vertical and diagonal lines. Its capability was tested by presenting it with a pattern after training and seeing if it recognised it. These processes were then automated, so that the network could be trained and tested many times, producing a series of results, from which graphs were produced. The amount of noise was automatically changed for training and testing, increasing from no noise at all to some suitable value; the maximum noise level being 100. Finally, the patterns were changed, to find out how successful the network is at recognising different patterns. Background and Theory This investigation is based on and artificial neural ...

See this essay or coursework document in full right now