Keywords
Forensic Science, Abstract Classification, Topic Modeling, Latent Dirichlet Allocation (LDA),Machine Learning, XGBoost,K-Nearest Neighbors (KNN), Multinomial Logistic Regression.
Abstract
Effective classification of scientific abstracts, especially in forensic science, is crucial for streamlined information access. Our study addresses the challenges of classifying unlabelled abstracts by leveraging Latent Dirichlet Allocation (LDA) to unveil hidden themes. The methodology refines these topics for classification using advanced models like XGBoost, KNN, and Logistic Regression. Beyond precision in abstract categorization, our approach aims to uncover hidden structures within forensic science literature, providing a nuanced understanding of its dynamics. The study explores related research, outlines methodology, analyzes results, and concludes that integrating LDA with XGBoost enhances abstract categorization, affirming proficiency in representing diverse forensic science topics.
Semester
2026
Course Name
STA 6366 Data Science 1
Instructor Name
Rui Xie
College
College of Sciences
STARS Citation
Alipour Yengejeh, Amir, "Enhancing Forensic Science Abstract Classification: A Topic Modeling Approach" (2026). Data Science and Data Mining. 54.
https://stars.library.ucf.edu/data-science-mining/54
Included in
Accessibility Statement
This item was created or digitized prior to April 24, 2027, or is a reproduction of legacy media created before that date. It is preserved in its original, unmodified state specifically for research, reference, or historical recordkeeping. In accordance with the ADA Title II Final Rule, the University Libraries provides accessible versions of archival materials upon request. To request an accommodation for this item, please submit an accessibility request form.