A Comparison of Data-Driven Automatic Syllabification Methods

Although automatic syllabification is an important component in several natural language tasks, little has been done to compare the results of data-driven methods on a wide range of languages. This article compares the results of five data-driven syllabification algorithms (Hidden Markov Support Vector Machines, IB1, Liang's algorithm, the Look Up Procedure, and Syllabification by Analogy) on nine European languages in order to determine which algorithm performs best over all. Findings show that all algorithms achieve a mean word accuracy across all lexicons of over 90%. However, Syllabification by Analogy performs better than the other algorithms tested with a mean word accuracy of 96.84% (standard deviation of 2.93) whereas Liang's algorithm, the standard for hyphenation (used in TEX), produces the second best results with a mean of 95.67% (standard deviation of 5.70).