Comparison of Tree-Based Machine Learning Algorithms to Predict Reporting Behavior of Electronic Billing Machines

Abstract

Tax fraud is a common problem for many tax administrations, costing billions of dollars. Different tax administrations have considered several options to optimize revenue; among them, there is the so-called electronic billing machine (EBM), which aims to monitor all business transactions and, as a result, boost value added tax (VAT) revenue and compliance. Most of the current research has focused on the impact of EBMs on VAT revenue collection and compliance rather than understanding how EBM reporting behavior influences future compliance. The essential contribution of this study is that it leverages both EBM’s historical reporting behavior and actual business characteristics to understand and predict the future reporting behavior of EBMs. Herein, tree-based machine learning algorithms such as decision trees, random forest, gradient boost, and XGBoost are utilized, tested, and compared for better performance. The results exhibit the robustness of the random forest model, among others, with an accuracy of 92.3%. This paper clearly presents our approach contribution with respect to existing approaches through well-defined research questions, analysis mechanisms, and constructive discussions. Once applied, we believe that our approach could ultimately help the tax-collecting agency conduct timely interventions on EBM compliance, which will help achieve the EBM objective of improving VAT compliance.

Prediction of Tuberculosis From Lung Tissue Images of Diversity Outbred Mice Using Jump Knowledge Based Cell Graph Neural Network

“…The introduction of the CG-JKNN model in our study presents new avenues for potential research. The performance of GNN models can be significantly enhanced through knowledge distillation [ 88 ], [ 89 ]. This process could enable the GNN models to require even fewer parameters than XGBoost while delivering comparable performance.…”

“…While this is the lowest among the models chosen, it still represents a decent level of performance. Feature scaling was omitted in our approach for ML models, as these models are tree-based and inherently robust to scaling [89]. In developing our GNN models, we employed the Standard Scaler technique [90] for feature scaling, as GNN models require scaled features to ensure that each input feature contributes proportionately to the model's learning process [40].…”

Challenges of Automated Identification of Access to Education and Training in Germany

“…We find some recent publications on transformer models [22][23][24]. In general, all methods have their own biases and assumptions that need to be carefully considered, see for example [25][26][27].…”

Digital taxation, artificial intelligence and Tax Administration 3.0: improving tax compliance behavior – a systematic literature review using textometry (2016–2023)

“…Recent research in digital taxation and AI reveals significant advancements in promoting tax compliance. Bellon et al (2022) and Murorunkwere et al (2023) both demonstrated how digital tools such as electronic invoicing and machine learning algorithms such as random forest effectively improve compliance among businesses and predict value added tax (VAT) compliance. These studies underscore the role of technology in reducing compliance costs and enhancing enforcement.…”