Audit Data Analytics part 2

Examples of predictive analytics

-Regression

-Classification

-Probability

-Sentiment analysis

• Good for AFDA and fair value

Regression

Predicts specific dependent values based on independent variable inputs

• Easiest one to think about

• If its a good fit then we can predict allowance

Classification

Predicts a category for a record, mainly focused on risk assessment. The predicted classes may be low risk or high risk.

• Are customers going to pay or are they not

• Takes time to build up audit

• Strength of the class can be important ex. 10 riskiest transactions

Probability

Uses a rank score to evaluate the strength of classification

• Risk based audit approach

Sentiment analysis

Evaluates text for positive or negative sentiment to predict positive or negative outcomes

• Used more by analysts than auditors

Prescriptive examples

Doesn’t really come in to external auditing but does in internal

• What if analysis

• Applied statistics

• AI

Example prescriptive audit procedure

Analysis determines procedures to follow when new accounts are opened for inactive customers, such as requiring approval

What if analysis

Decision support systems

• what impact will these things have

Applied statistics

Predicts a specific outcome

• how do we get to out goal, what are the variables

AI

Uses observations of past actions to predict future actions for similar events. Models expected behavior by evaluating past actions taken by auditors to predict expected behavior in an unknown case.

• how do we takeover this market

• big ticket things

Predictive and prescriptive analytics provide less _________ output and more _________

models

deterministic, probabalistic