Page 34 - 臺大管理論叢第33卷第1期
P. 34

An Integrated Data-Driven Methodology for Auditor Performance Appraisals and Auditor Assignment
               Optimization



               An Integrated Data-Driven Methodology for Auditor
               Performance Appraisals and Auditor Assignment Optimization


               Tzu-Chien Wang, Department of Business Administration, National Taiwan University



                                           1. Purpose/Objective


                    With the expansion of related business groups and the impact of digitization on

               financial industry in Taiwan, in recent years, financial holding companies often fail
               in internal supervision and management due to problems of scheduling auditing-staff.
               Although enterprises and corporate management organizations generally believe that
               the large amount of data collected from the performance management system may have
               great practical value for the audit departments, most of the existing studies only focus

               on verifying the relevance between auditing performance and variables. Conversely,
               our study emphasizes on the field application of financial control audit. Summarizing
               from previously established models of audit performance evaluation, we discuss and

               further establish the most optimal performance assessment model using machine learning
               regression. Once identify some key factors, we use an integer planning method to
               construct an optimal auditor assignment model to assist financial holding companies in
               analyzing the existing audit workload and provide a better audit assignment mechanism to
               facilitate human resource decisions.



                                   2. Design/Methodology/Approach


                     To accurately estimate the audit performance of different auditors under different

               tasks and to understand the best personnel assignment approach under different data
               constraints, using audit data of financial holding companies in Taiwan, we integrate the
               random forest technology and heuristic algorithm to construct an analysis model, which
               could apply to evaluate the audit performance of internal auditors and optimally assign

               audit tasks according to the required performance value.
                    Since these companies do not provide all the performance data of the assigned
               personnel, to generate complete audit data for the assignment planning of senior



                                                      26
   29   30   31   32   33   34   35   36   37   38   39