IMPLEMENTATION OF THE CONTEXTUAL BAYESIAN MODEL FOR PERSONALIZED, REASONED MULTI-CRITERIA SCHOOL ORIENTATION: APPLICATION TO THE IVORIAN CONTEXT

School orientation in sub-Saharan Africa reflects structural inequalities linked to geographic origin, socio-occupational category, and gender, an observation that motivated the design of a contextual Bayesian system recommending a post-baccalaureate track based on the student's academic profile, psychological profile, and socio-economic situation. This article examines the conditions for its real-world deployment in Côte d'Ivoire, beyond its algorithmic performance alone. An offline-capable technical architecture is proposed for institutions lacking reliable internet access, combined with a progressive deployment that respects local institutional cycles. The contextual variables are redefined for the Ivorian context, with the addition of a gender variable correcting orientation biases documented in the literature. A multidisciplinary governance framework monitors algorithmic bias and guarantees the transparency of recommendations to families. A change-management strategy enriches the guidance counselor's role rather than replacing it, through a usage protocol based on the confidence level of the recommendations. High algorithmic performance thus remains a necessary but not sufficient condition for a real impact on school equity.