Figure 6
The flow diagram is arranged from left to right. On the left, a rounded rectangle labeled “Machine Learning in Fraud Detection” serves as the starting node. From this box, a vertical branching line extends rightward and splits into four horizontal branches leading to four category boxes arranged from top to bottom: “M L Techniques”, “Specific M L Methods”, “Application Areas”, and “Data Sources and Analysis”. From “M L Techniques”, rightward branches list five items: “Supervised Learning”, “Unsupervised Learning”, “Deep Learning”, “Natural Language Processing (N L P)”, and “Data Mining”. From “Specific M L Methods”, rightward branches list seven items: “Classification Algorithms”, “Anomaly Detection”, “Text Mining and Analytics”, “Topic Modeling”, “Sentiment Analysis”, “Support Vector Machines (S V M)”, and “Artificial Neural Networks (A N N)”. From “Application Areas”, rightward branches list four items: “Financial Fraud Detection”, “Credit Card Fraud Detection”, “Insurance Fraud”, and “Money Laundering”. From “Data Sources and Analysis”, rightward branches list three items: “Textual Analysis”, “Financial Analysis”, and “SustA Inability Reporting”.Machine learning cluster conceptual map. Source: The Authors