Sentilyze: A Bilingual Sentiment Analysis API Using Logistic Regression and Support Vector Machine Algorithms

##plugins.themes.bootstrap3.article.main##

Juan Gabriel Baterina Jared Castañeda Stephen Grant Sumadsad

Abstract

This study introduces a Bilingual Sentiment Analysis API employing Logistic Regression and Support Vector Machine (SVM) algorithms. Using descriptive research methods, the study focuses on the administrative personnel of the External Relations Department (ERD) at Colegio de San Juan de Letran Calamba, with the Director as the primary respondent. Data were collected through interviews and social media content, particularly Facebook posts and comments. The API development followed the Agile Model within the Software Development Life Cycle (SDLC). Results highlight the proposed model’s precision and reliability, with Logistic Regression delivering accuracy and SVM effectively capturing sentiment patterns. Combined, these algorithms achieved a 90.19% accuracy rate across five recorded tests, showcasing the efficiency of automated sentiment analysis in reducing processing time and providing real-time insights. This tool aids the ERD in online reputation management and strategic decision-making by interpreting sentiment data. Furthermore, the research advances natural language processing by addressing sentiment analysis challenges in social media contexts and offering a practical solution for sentiment interpretation. The Bilingual Sentiment Analysis API provides a powerful resource for improving the ERD's capabilities in utilizing sentiment data to enhance communication strategies and operational efficiency.

##plugins.themes.bootstrap3.article.details##

Section
Articles