Control risk assessment in auditing: Artificial neural network approach

The purpose of the study is to develop a decision support tool with Artificial Neural Network for the use of external auditors in assessing preliminary control risk level. A survey was conducted to collect internal control data of top 500 industrial corporations in Turkey. The number of surveys returned was 169 out of 500. An external auditor assessed control risk level for each corporation. While the research used 100 firms to develop an artificial neural network model, the remaining 69 firms were selected for testing the model. The findings indicated that the artificial neural network model could mimic the auditor’s decision. The category accuracy rate between the predictions of the artificial neural network model and the auditor opinion was 71.01%. In conclusion, the artificial neural network model provided a useful decision support tool for auditors to assess preliminary control risk level.