AI Based Predictive Analytics Model for Intelligent Decision Support in Digital Systems

Abstract

Many organizations still rely on the old paradigm of relying on reports that are not real-time, and manual judgment when it comes to critical decisions. This paper presents an intelligent decision making approach based on predictive analytics model that predicts the behavior of digital systems by the Explainable XGBoost based approach developed in Google Colab. The model takes data preprocessing, feature engineering, supervised learning, hyperparameter tuning, performance evaluation and explanation using SHAP into consideration and transforms the historical and current digital indicators into trustworthy decision insights. The proposed framework can predict system properties like risk, performance degradation, failure probability or decision priority, and can explain the key factors that affect each of these properties. A prototype evaluation demonstrates that Explainable XGBoost model outperforms other baseline machine learning models in terms of accuracy, precision, recall, F1 score and ROC AUC. The study illustrates how to achieve the goals of transparency, trust and effective decision making in today's digital governance environments by leveraging predictive analysis and interpretability.