Student Orientation Piloted by Artificial Intelligence Using Filtering Collaborative Base on Memory
Camile LIKOTELO BINENE *
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
Pierre J. SAKODI MJANAHERI
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
Luz MPEMBA NGOMA
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
Guylit KIALA LUTUMBA
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
Pierre KAFUNDA KATALAY
Faculty of Science and Technology, Department of Mathematics and Computer Science, University of Kinshasa (UNKIN), Democratic Republic of Congo.
Cédric KABEYA TSHISEBA
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
Franci MAYALA LEMBA
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
Boniface ENGOMBE WEDI
Faculty of Science and Technology, Department of Mathematics and Computer Science, University Educational National (UPN), Democratic Republic of Congo.
*Author to whom correspondence should be addressed.
Abstract
Traditional student orientation in universities and higher education institutions in the Democratic Republic of Congo, particularly at the National Pedagogical University, often depends on personal choice, social influence, perceived employment opportunities and ease of enrolment rather than on candidates' demonstrated academic skills. This practice may contribute to academic failure, uneven distribution of students across departments and weak alignment between graduates' training and labour-market expectations. This study proposes a memory-based collaborative filtering algorithm for recommending university courses according to students' prior academic performance. The model uses two main matrices: one containing candidates' average marks in subjects studied at secondary-school level and another containing the weighting of those subjects in university programmes. The product of these matrices provides a compatibility structure that supports comparison between candidate profiles and programme requirements. Cosine similarity is then used to identify the programme with the strongest alignment for each candidate. The algorithm was implemented in Python using Jupyter Notebook within the Anaconda environment and tested on a sample drawn from the Faculty of Sciences. The results show candidate placement across Biology, Geography, Petrochemistry, Physics, Mathematics and Statistics, and Hospitality, with recommendation scores indicating the relative strength of alignment. The proposed approach offers a structured decision-support tool for academic guidance while retaining the need for institutional interpretation and validation.
Keywords: Recommendation system, collaborative filtering, student-candidate orientation, artificial intelligence, algorithm, machine learning, cosine similarity, prediction, matrix, academic guidance