Tagging Experiments Using Neural Networks
Martin Eineborg, Björn Gambäck · DSpace repository (University of Tartu) · 1994
The paper outlines a method for automatic part-of-speech tagging using artificial neural networks.Several experiments have been carried out where the performance of different network architectures have been compared to each other on two tasks: classification by overall part-of-speech (noun, adjective or verb) and by a set of 13 possible output categories.The best classification rates were 93.6% for the simple and 96.4% for the complex task.These results are rather promising and the paper compares them to the performance reported by other methods; a comparison that shows the neural network completely compatible with pure statistical approaches.