Use of the WordFrame Model in Learning the Filipino Morphological Process

This paper presents the results of the adoption of the WordFrame model to the Filipino language. The WordFrame model, developed by Richard Wicentowski, is a robust multi-lingual supervised learning algorithm that learns a language’s morphological process. This algorithm models a word as a seven-way split, namely, the canonical prefix, the point-ofprefixation change, common substrings, vowel change, point-of-suffixation, and canonical suffix. The problem with this division is that it cannot correctly model infixes and reduplicating stem in a word, which exist in the Filipino language. For instance “kumakain” includes the infix “um” and the reduplicated stem “ka”. The algorithm was modified to correctly model these two morphological phenomena. Results show an accuracy rate of 89% in root word determination. Although this is significantly lower than the 97% accuracy rate shown by Wicentowski’s research, his work was only tested on Tagalog verbs, while this research included words in the other parts-of-speech.