Processed Synthetic Taxpayer Risk Identification Dataset and Machine Learning Analysis Code

This repository contains the processed synthetic taxpayer risk dataset and Python code used in the study “Applying Machine Learning methods to Detect and Cluster Taxpayer Risk Profiles.” The dataset was derived from the publicly available kaggle dataset “Tax Risk Identification Dataset” by Ziya, available at https://www.kaggle.com/datasets/ziya07/tax-risk-identification-dataset. The deposited files include the original dataset, a table describing all variables, theoretical and conceptual framework diagrams created from drawio.com, feature engineering (data generation and preprocessing), machine learning classification and clustering, model evaluation, figures generation code reported in the manuscript. The data and code are provided to support transparency, reuse, and replication of the reported taxpayer risk profiles classification and clustering analyses. The dataset is synthetic and does not contain personally identifiable taxpayer information.