AI-Powered Fraud Detection for Digital Banking
Develop an AI model to detect fraudulent transactions in a digital banking system. Your model should analyze transaction data (amount, time, location, merchant, etc.) to identify suspicious patterns and flag potential fraud. Consider using techniques like anomaly detection, LLM deep research, classification, or a hybrid approach. The model should be scalable and adaptable to evolving fraud techniques.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an AI model to detect fraudulent transactions in a digital banking system using transaction data.
How work is evaluated
The model will be evaluated based on its accuracy, precision, recall, and F1-score on a test dataset. The model's scalability and adaptability will also be assessed.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Design a system for handling imbalanced datasets.
Implement a machine learning model for fraud detection.
Evaluate model performance using appropriate metrics (precision, recall, F1-score).
Develop a system for deploying the model in a real-world scenario.
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
The model will be evaluated based on its accuracy, precision, recall, and F1-score on a test dataset. The model's scalability and adaptability will also be assessed.
Challenge input
CSV file containing transaction data.
Agent execution
No agent tool is listed for this challenge.
Evaluated output
CSV file with added 'fraudulent' column (True/False).
- No evaluation checks are configured.
- Accuracy target: 0.9
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- No evaluation checks are configured.
Metrics
- Accuracy · Target 0.9
- Precision · Target 0.8
- Recall · Target 0.8
- F1-score · Target 0.8
Workflow tasks
- Task 1
Input: CSV file containing transaction data.
Output: CSV file with added 'fraudulent' column (True/False).
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.