Michael Dimitrov
E-mail: m.m.dimitrov@rndc.bg
Michael Dimitrov is an associate professor (PhD in "National Security") in Rakovski National Defense College in Sofia. His scientific interests are in the field of energy security, formalized representation of regional spaces and automated processing of data.
pp. 9-22
https://doi.org/10.70265/HCQX8107
This study compares the performance of two web content extraction tools, one based on computer vision (EE) and the other implementing Puppeteer (3P-IO), in the context of their ability to quantify the regional security. Using a dataset of 100 webpages, the evaluation focuses on crawl time, scrape time, precision (BLEU score), and load success rate. The results reveal significant differences: 3P-IO is faster in both crawl and scrape times, while EE achieves a 100% load success rate, ensuring reliable content retrieval. Despite similar precision, EE’s reliability is a key advantage. The study proposes a hybrid approach that combines 3P-IO’s speed and precision with EE’s reliability, activating the computer vision component only when 3P-IO’s fallback strategy fails. This approach offers a balanced solution, leveraging the strengths of both tools. The study provides valuable insights for stakeholders interested in the measurability of regional security dynamics.
Web Content Extraction, Regional Security, Computer Vision, Headless Browsing, Scrape Time, Precision (BLEU Score), Hybrid Approach.
Dimitrov, M. (June 2025). OSINT Tools for Regional Dynamics Evaluation. Security and Defense(1), 9-22. https://doi.org/10.70265/HCQX8107
Dimitrov, Michael. "OSINT Tools for Regional Dynamics Evaluation." Security and Defense (Vasil Levski National Military University Publishing Complex), № 1 (June 2025): 9-22. https://doi.org/10.70265/HCQX8107
Dimitrov, M., 2025. OSINT Tools for Regional Dynamics Evaluation. Security and Defense, June, Issue 1, pp. 9-22. https://doi.org/10.70265/HCQX8107
Dimitrov, Michael. "OSINT Tools for Regional Dynamics Evaluation." Security and Defense 1 (2025): 9-22. https://doi.org/10.70265/HCQX8107.