Bezza ; Asma2026-09-272026-09-272026http://dspace.univ-khenchela.dz:4000/handle/123456789/11375This thesis presents DZ-UniGuide AI, an intelligent academic orientation system designed to assist new Algerian bachelors in selecting suitable university programs. The study begins withthecollectionandanalysisofrealstudentdata,consistingof907responsesthatcapture academic performance and personal interests. In parallel, a reference database of 1,717 academic programs was build from official Ministry documents, including admission criteria and field descriptions. Basedonthisdata,aMachineLearning(ML)modelusingRandomForestwasdeveloped to predict students’ probability of success, achieving an accuracy of 90.7%. To better understand students’ preferences, a Natural Language Processing (NLP) module based on TF-IDF and cosine similarity was implemented to analyze free-text inputs describing their interests. The system also integrates key constraints specific to the Algerian higher education system, including academic eligibility calculated through the official weighted average and geographic eligibility determined by the student’s baccalaureate Wilaya. A bilingual (Arabic/English) web application was developed and deployed on Streamlit Cloud to provide an accessible and interactive user experience. The system then was evaluated using quantitative metrics (precision: 84%, recall: 99%, F1-score: 0.91) as well as user feedback collected through a survey (N = 21). The results show an overall satisfaction rate of 77%, with 85% of users reporting that the interface is easy to use. This work demonstrates that combining ML and NLP effectively creates a personalized, transparent, and context-adapted orientation tool for the Algerian higher education system.enThesis