Intelligent Auditing System Based on Deep Belief Network Density based Spatial Clustering of Applications

In recent years, as the traditional manual audit systems disposed to errors and frequently focused on compliance, rather than risk valuation the need of advanced audit systems had increased. Additionally, increased complex financial transactions, advanced technologies and regulators requirements in sectors and industry firms have necessitated the developing intelligent audit systems. The existing methods for developing intelligent audit systems failed in providing valuable indicators for audit process. Therefore, this paper proposes Deep Belief Network (DBN) - Density Based Spatial Clustering of Applications with Noise (DBSCAN) for designing an intelligent audit system by detecting fraudulent firms. Initially, audit data dataset is taken which incorporated audit data information of various firms. The collected audit data is preprocessed by utilizing data cleaning and feature engineering, for obtaining efficient financial audit data. Then this financial audit data features are extracted by employed DBN method for revealing hidden patterns. The extracted relevant features are evaluated for detecting the fraudulent firms by utilizing DBSCAN data mining method. The proposed DBN-DBSCAN achieved better results which comprises accuracy (96.58%), Specificity (97.85%), Sensitivity (95.69%), and F1-score $\mathbf{(9 5. 2 0 \%)}$ when compared with existing Long Short-Term Memory (LSTM).