A neural network that learns to do hyphenation

摘要

Summary form only given, as follows. Hyphenation of German wordsis a highly irregular problem. Existing solutions for automatichyphenation are not very satisfying. A `sequential network' was appliedto this problem. The training algorithm was standard backpropagation.The network was trained with a collection of 1000 German words togetherwith their correct hyphenation. In subsequent tests with unknown words,a correctness of 96.8 percent was achieved. Analysis of the simulationresults indicates that with further increases of the training dataimprovements are still possible