A Morphological Analyzer for Filipino Verbs

A Morphological Analyzer for Filipino Verbs* Robert R. Roxasa and Gersam T. Mulaa aUniversity of the Philippines Visayas – Cebu College Cebu City, Philippines rrroxas@upv.edu.ph and gersammula@gmail.com Abstract. This paper presents a morphological analyzer that accepts Filipino verbs conjugated in different forms as inputs and analyzes them to produce the affixes used, the infinitive forms, and the tenses of the original input verbs. A prototype system was implemented and was fed with a file containing 1,050 Filipino verbs conjugated in various tenses using different types of affixes. The preliminary result showed that the accuracy rate was high in three expected outputs, i.e., tenses, infinitive forms, and affixes used. Keywords: morphological analyzer, context-driven machine translation system, Filipino verbs 1.Introduction Filipino is a language whose verb conjugations are very complex because it uses different affix (prefix, infix, and suffix) combinations and even duplication of syllables or words. When looking up a word in Filipino dictionary, one needs to know the infinitive form of the verb to be able to find it. This is a difficult task for someone who is not familiar with the way Filipino verbs are conjugated into any of the three tenses: Pangnagdaan (Past), Pangkasalukuyan (Present), and Panghinaharap (Future). These three are commonly known today as Aspektong Perpektibo, Aspektong Imperpektibo, and Aspektong Kontemplatibo, respectively (Dizon, 2006). Because of the complexity of the language, automated morphological analyzers will be difficult to construct. So researches should be focused on how to capture all possible forms so that a machine translation system can produce accurate translation. One approach is to use a lexicon that stores the context-words that help determine the appropriate equivalent word(s) in the target language. The lexicon must use the infinitive forms of the verbs to facilitate the look up of verbs and use headwords for non-verbs. So with this kind of lexicon in mind, a morphological analyzer that returns the infinitive form of a verb, its tense, and its affix(es) used is necessary. 2.Review of Related Works When someone attempts to create any machine translation system for natural languages, it is not possible to do away with morphological analyzers. One area of focus for morphological analyzers is the analysis of verbs. Verbs are usually conjugated according to tense, number, voice, and mode while nouns and adjectives are usually declined according to person, number, case, and degree. Some languages have simple forms of verb inflection or conjugation while other languages have complex forms of inflection. Japanese was originally thought to have simple verb inflection, thus it was not the central subject on Natural Language Processing * Copyright 2008 by Robert R. Roxas and Gersam T. Mula 467 22nd Pacific Asia Conference on Language, Information and Computation, pages 467–473 (NLP). But in recent past, it had become an important subject (Hisamitsu and Nitta, 1994). For other languages, verbs are also given much attention in research. These languages include Chinese (Kim, et. al, 2002), Japanese (Nakamura, 2007), and Korean (Hong, et. al., 2004; Jun, 2007) to name a few. For languages …

Content truncated to 500 words