AI-POWERED TAX FRAUD DETECTION SYSTEM FOR INCOME TAX FILINGS

The detection system uses AI and the Isolation Forest method implemented through Apache Spark to examine extensive financial data in order to detect tax fraud. The proposed detection system studies unusual tax filing behaviors because criminal tax payers behave differently from usual financial patterns. Isolation Forest acts as an unsupervised machine learning technique because it serves as an outlier detection method to identify questionable tax records while needing no dependent labeled data. The implementation of Apache Spark brings the ability to handle enormous financial transaction data alongside the provision of efficient scalable operations across numerous sources. This synthetic tax data methodology performed successful detection of suspicious cases without creating many wrong identifications. Small tax conduct nuances serve as the basis for the system to provide active oversight capabilities to tax authorities. Department of Revenue uses digital governance to achieve better compliance monitoring through data intelligence as well as enhanced system tracking capabilities Keywords: Tax Fraud Detection, Isolation Forest, Apache Spark, Anomaly Detection, Income Tax Filings, Machine Learning, Big Data Analytics. Type of paper: Research Paper