Hybrid K-Nearest Neighbors with Ant Colony Optimization for Securing data warehouses against inferences

Abstract

Data Warehouses (DWs) are among the most powerful technologies for storing and managing large volumes of corporate data, which often include sensitive or confidential information. However, they remain vulnerable to inference attacks and insufficient access control mechanisms. In recent years, Artificial Intelligence (AI) techniques have become key tools for enhancing DW security, particularly for detecting and preventing unauthorized inferences. In this work, we propose a hybrid approach that combines the K-Nearest Neighbors (KNN) classification algorithm with the Ant Colony Optimization (ACO) metaheuristic to strengthen data warehouse security. The objective is to minimize inference risks by optimizing variable selection and improving the accuracy of sensitive data classification. The proposed ACO–KNN model was evaluated using a dataset of 1,000 SQL analytical queries generated by the IBM Db2 Query Manager, representing realistic decision-support workloads. Experimental results show that the hybrid model significantly outperforms traditional KNN and other metaheuristic-based methods in terms of prediction accuracy, convergence speed, and inference prevention capability. This demonstrates the model’s potential for practical integration into Business Intelligence (BI) and OLAP environments, contributing to more secure and reliable analytical decision-making.