Natural language grammatical inference with support vector machines

Abstract

Grammatical inference refers to the process of learning of grammars and languages from data. Machine learning of grammars finds a variety of applications in syntactic pattern recognition and natural language acquisition. Grammatical inference has become one of the key methods for organizing online information, since hard-coding the classification rules is costly or even unpractical. The paper presents a new approach to perform grammatical inference using support vector machines (SVM). SVMs are a class of algorithms that combine the principles of statistical learning theory with the optimisation techniques and the idea of a kernel mapping. The paper considers replacing the inference algorithm with support vector machines and the grammar is that of English language. The accuracy is found to be 97.8 % which is much better than conventional machine learning methods.