Automated Data Extraction Using Web Scraping Techniques and AI for Creating News and Events Database in Thailand's Southern Border Provinces

Abstract

   This research aims to develop and apply an automated data extraction system utilizing Web Scraping techniques and Artificial Intelligence (AI) to establish a comprehensive database of news and events in Thailand's Southern Border Provinces. Furthermore, the study evaluates the system's technical and quantitative performance to enhance information management efficiency at the John F. Kennedy Library, Prince of Songkla University, Pattani Campus.
   The implementation process consists of four main stages: (1) data collection via Web Scraping from over 15 targeted national and local news sources; (2) data processing and classification using the Naive Bayes Classifier algorithm; (3) data integration and redundancy reduction; and (4) database design and storage.
      The findings indicate that the system significantly reduced data archiving time from 1–2 hours to merely 10–15 minutes per day, representing a reduction of approximately 85%. The system demonstrated a high extraction accuracy of 97% and a duplicate detection rate of 98%. Additionally, the classification accuracy reached 92%, while the processing error rate was reduced to only 2%. The system successfully expanded coverage to include both national and local levels. Expert evaluation rated the system’s overall efficiency as "very good" ( = 4.85, SD = 0.18), particularly highlighting its precision and processing speed. These results significantly contribute to supporting policy planning, situational monitoring, and strategic decision-making in the region, while establishing a reliable database foundation for sustainable future application.