Auditing & Assurance › Analytical Procedures
How do you use regression or predictive models in substantive analytics?
AdvancedAnswer
Regression models predict an account (e.g., utility cost) from drivers (production volume, degree-days, prices), producing a statistically-based expectation with a measurable precision/threshold. Recorded amounts outside the predicted range are investigated. Such models increase precision and objectivity but require reliable driver data and validation of the relationship.
Real-world example A regression of energy cost on production and temperature flags a month whose actual cost falls well outside the predicted band.