Abstract
This study aims to analyse the detection of corruption by Auditor within the implementation of a corruption detection device, which called as HU-Model. This corruption detection device was built from the components which can cause a corruption, such as fraud star with components of pressure, opportunity, ability, justification, and loss of integrity. The first stage of implementing HU-Model was carried out by distributing questionnaires to the work units on the Ministry of Educational Culture Research and Technology of Indonesia by the total of 165 respondents begin from echelon unit 1 at the centre, unit 2 and unit 3 of a region and several state universities in Indonesia. Furthermore, the corruption detection with HU-Model is carried out by collecting secondary data in form of 52 formulas with more than 104 raw data on 15 sample units. Secondary data collection and running data carried out using an online application built by the team based on HU-Model. Data analysis was carried out using Classification and Regression Tree (CART) method program to make a prediction model for the classification of corruption detection. The data analysis model used is an econometric model with an analysis technique of the least squares model. The software used in conducting data analysis is STATCAL. The implementation indicated that there are 147 units in clean (green) cluster, 17 units in grey cluster, and a unit indicated the red cluster (corruption). This HU-Model is quite effective to assist auditors in detecting corruption, determining audit scope and useful for the effort to prevent corruption.
Keywords: HU-Model, fraud detection system, auditory, corruption detection, corruption prevention
JEL Classification: M41, M42, M48