Abstract : This research investigates the use of explainable machine learning models for stock price prediction in the Nigerian financial market by the use of long-term historical data from GTCO (1996-2024), Access Holdings (1998-2024), and Zenith Bank (2004-2024). Two models, Random Forest and XGBoost were developed and compared, and SHAP (SHapley Additive exPlanations) values were used to interpret the impact of each predictor. This methodological approach melds prediction accuracy with explainability to not only quantify the performance but also highlight the important features. The findings indicate that XGBoost recorded lower prediction errors than Random Forest, thereby showing a greater capacity to capture the complex and unstable non-linear nature of Nigerian banking stocks. SHAP analysis showed that the moving averages, recent returns, volatility, and trading volume were the leading predictors. Moreover, the banks illustrated varying degrees of sensitivity to market signals: GTCO’s predictions were influenced mostly by trading volume, Access Holdings was more responsive to the short-term trends, whereas Zenith Bank was dependent on the broader market indicators. The research offers a transparent system that not only raises the performance of the prediction but also the interpretability. It serves as a scarce resource of explainable AI in the African financial markets and thus, provides the real-world implementable insights for investors, analysts, and policymakers.

Keyword : Explainable Machine Learning, Stock Market Prediction, and SHAP Values, XGBoost.