THE INFLUENCE OF SECURE DATA SYSTEMS ON FRAUD DETECTION IN BUSINESS INTELLIGENCE APPLICATIONS

Abstract

This study investigated the influence of secure data systems on fraud detection performance within business intelligence (BI) applications through a quantitative analysis. The empirical phase examined five core security dimensions—access control maturity, encryption coverage, monitoring completeness, pipeline security, and data governance strength—across 220 organizations using BI-driven fraud detection. Descriptive analysis showed moderate to high implementation of security features, with mean scores ranging from 4.8 to 5.4 on a seven-point scale. Fraud detection performance indicators demonstrated substantial variability, with detection accuracy averaging 78.4%, false-positive rates averaging 22.7%, and time-to-detection ranging from 1 to 72 hours. Correlation analysis revealed strong positive associations between monitoring completeness and detection accuracy (r = .52), as well as between pipeline security and output consistency (r = .54). Regression analysis indicated that secure data system features explained 39% additional variance in detection accuracy beyond organizational controls. Monitoring completeness (β = .28), pipeline security (β = .24), and governance strength (β = .22) emerged as the strongest predictors. Access control maturity demonstrated a significant negative effect on false-positive rates (β = –.26), while encryption coverage showed meaningful influence on faster time-to-detection (β = .19). Mediation analysis found that security culture partially strengthened relationships between key security features and detection outcomes, whereas BI integration complexity weakened them. Overall, the findings demonstrated that secure data systems play a decisive role in enhancing fraud detection accuracy, stability, and responsiveness within BI environments, highlighting their importance as both protective and performance-enabling components of organizational analytics.