Leveraging Grobid For Bibliographic Metadata Extraction In The Impact Factor Assessment Of National Journals

Abstract

This study aims to evaluate the performance of GROBID in extracting bibliographic metadata and to use the extracted data for calculating journal impact factors. Based on the theoretical foundations of GROBID and the concept of impact factor, the research applies a full-text extraction method using GROBID to obtain metadata from scientific journal articles, which is then parsed and stored in a database for further analysis. The performance of GROBID is assessed by calculating the Mean Squared Error (MSE) from 30 randomly selected data points, yielding an MSE value of 0.0019, indicating minimal deviation between actual and predicted values. The findings reveal that Journal ID 151 (IAES International Journal of Artificial Intelligence/IJ-AI) has the highest impact factor at 0.5605, derived from a ratio of 250 citations to 446 publications, followed by Journal ID 147 at 0.5254, while Journal ID 18261 registers the lowest at 0.0580. These results highlight significant variations in influence and productivity among journals and confirm that GROBID is a reliable tool for metadata extraction used in impact factor calculations.