Author: JOHN C. VALDORIA

Issue: 2022-2023

Abstract:

This research emphasize on the translation endangered native Ayta Tayabas dialect through the implementation of natural language process through Naïve Bayes Algorithm. Android based smartphones were used as a communication medium for converting Ayta Tayabas dialect through use of Speech-to-Text API and direct text to preferred English-Filipino language while implementing Naïve Bayes Algorithm is utilized to recognize each translated word. The study provides awareness and a new way on integrating Natural Language Processing algorithms, notably implementation Naive Bayes Algorithm to the dataset of Ayta Tayabas dialect retrieved from the Bokabularyong Tayabasin archive, the Lexicographic study of Tayabas-Tagalog that undergone manual data retrieval, and verified with the linguistics experts; cleaning, mined and trained using python and implement Naïve Bayes Algorithm through spaCY API. Anaconda API were proven to be used as a validation tool for trained models based on the relevant studies done by other research study. The results were trained using classification models that could identify the Ayta Tayabas dialect at certain rate and accuracy based on the lexicon dataset integrated on the application. Therefore, Naïve Bayes classification exemplifies the translation of Ayta Tayabas dialect with corresponding English-Filipino language.

Keywords: Natural Language Processing, Naïve Bayes Algorithm, Ayta Tayabas, Name Entity Tagging, Language Translator