Proceedings of The 4th International Conference on Advanced Research in Social Sciences
Referencing Scientific Articles by LDA Technology
Boussaadi Smail, Hassina Aliane, Pr Abdeldjalil Ouahabi
Finding relevant scientific articles is a laborious process, which requires a lot of time and effort, especially on online journal databases, where article research is based on metadata such as keywords or authors’ names. In this context we are interested in the referencing of articles by topics they address, thanks to topic modeling technology as LDA applied, on article metadata (Title, abstract, keywords). Based On the preliminary results, it is shown that the combination of some metadata generates a distribution of words on topics of high coherence and at the same time allows to identify the subjects of interest of each scientific article with a better interpretability
keywords: Latent Dirichlet Allocation, Scientific Article, Topic Modeling.