Dina Mohamed Kamal Atito creator preparation. Ayman Ramadan Elkilany thesis advisor. Hoda Mokhtar Omar Mokhtar thesis advisor. text theses ua 2022 monographic eng ara 65 pages : illustrations ; 30 cm+ CD Recommendation systems are algorithms that aim to predict the users' needs and automatically suggest the most relevant items to the users. Recommender systems are becoming increasingly popular in our daily lives and applied in different domains to facilitate finding relevant and interesting items to the users. In the academic domain, the academic article recommendation systems have gained a lot of interest as an effective tool to suggest relevant articles for researchers according to their interests. An explicit identification of the topics of interest from the contents of academic articles that the researchers have authored, downloaded, or read has been always a challenging task. Accurate articles recommendation relies on the true identification of researchers{u2018} interests which is affected by the quality of the article's textual representation. In this thesis, we aim to improve the results of the academic recommendation system by enhancing the representation of the article and consequently enhancing the quality of the recommendation. In order to improve the representation of the articles, we focus on the semantic approaches to represent the words' semantic meanings rather than their syntactic representation only. In this thesis, two semantic representation models are proposed for articles representation, both models have been applied in the academic articles recommendation process specialized Dina Mohamed Kamal Atito ; Supervised Hoda Mokhtar Omar Mokhtar , Ayman Ramadan Elkilany. Thesis (M.Sc.) - Cairo University - Faculty of Computers and Artificial intelligence - Department of Information Systems Bibliography: pages 69-75. Issued also as CD Text in English and abstract in Arabic & English. LDA Latent Dirichlet Allocation(LDA) Recommendation Systems Word2vec 600 EG-GiCUC 220221 20260513123105.0 eng