Development Of a Machine-Learning Approach for Internal-Control Weakness Prediction and Audit-Risk Scoring In U.S. State Government Agencies

This study has investigated a machine-learning approach for predicting internal-control weaknesses and supporting audit-risk scoring in U.S. state government agencies. The study has examined whether machine-learning capability, internal-control data quality, organizational audit readiness, and auditor analytics competency influence internal-control weakness prediction effectiveness and audit-risk scoring effectiveness. A quantitative, cross-sectional, case-study-based design has been used with 286 valid responses from public-sector professionals, yielding an 86.67% response rate from 330 distributed questionnaires. The five-point Likert-scale data have been analyzed using descriptive statistics, Cronbach’s alpha, Pearson correlation, and multiple regression in SPSS. Reliability has been acceptable across all constructs, with alpha values ranging from .79 to .88 and an overall alpha of .90. Internal-control data quality has recorded the highest mean score, M = 4.14, followed by weakness prediction effectiveness, M = 4.08, audit-risk scoring effectiveness, M = 4.05, machine-learning capability, M = 4.02, auditor analytics competency, M = 3.96, and organizational audit readiness, M = 3.91. Data quality has shown the strongest relationship with weakness prediction effectiveness, r = .69, while weakness prediction effectiveness has had the strongest relationship with audit-risk scoring effectiveness, r = .72. The first regression model has explained 61.2% of the variance in weakness prediction effectiveness, with data quality as the strongest predictor, β = .33. The second model has explained 58.4% of the variance in audit-risk scoring effectiveness, with weakness prediction effectiveness, β = .47, and auditor analytics competency, β = .36, as significant predictors. All five hypotheses have been supported. The findings have indicated that reliable data, analytical capability, organizational readiness, and auditor competency are essential for transparent, evidence-based audit-risk scoring and stronger public accountability.