The loan prediction system tackles the difficulties encountered by many financial institutions in efficiently evaluating and assessing a large amount of loan applications. Conventional approaches often involve time-consuming manual assessments which lead to a large backlog of pending applications causing delays and occasional bias, resulting in inconsistent decisions.
Our solution to this problem is to improve the process by using artificial intelligence. This involves using AI to predict the likelihood of loan approval based on previous customer behaviour and historical data.
Automate the loan approval process to improve efficiency and reduce manual workload. Ensure unbiased lending practices by using historical data to ensure objectivity. Evaluate and minimise the risk of financial losses for financial institutions by accurately predicting the likelihood of a customer not being able to repay the loan. Provide clear and interpretable explanations for loan approval decisions to both customers and bank employees (Explainable AI) Use machine learning/deep learning algorithms to improve the precision of loan approval predictions. Minimise false positives and false negatives. The system should be scalable in order to handle large volume users and loan applications efficiently. Enhance customer experience by providing quick and unbiased loan decisions.
Data verification (Loan Approval / Rejection) Login and Sign up for customer and bank staff Data (personal and financial information) submission for application Perform Model Training View Model Performance Update and deploy model Authentication when submitting application
- Scenario 1 - Submitting Application
- 1.1. As a loan applicant, I want to be able to submit personal and financial details through the app for the AI system to assess eligibility.
- Scenario 2 - Checking Application Status
- 1.2. As a loan applicant, I want to be able to view the status of the loan application in real-time on the information page to see if it is approved or denied.
- Scenario 3 - Handling Account
- 1.3. As a loan applicant, I want to create a new account to submit the loan application through the system.
- 1.4. As a loan applicant, I want the system to securely handle and protect my personal data.
- 1.5. As a loan applicant. I want to be able to edit my personal details so that it stays up to date.
- Scenario 1: Verifying Applicant Information
- 2.1. As a bank employee, I want to be able to access the application to verify the accuracy of the information provided by the applicants.
- 2.2. As a bank employee, I want a user-friendly interface that provides an overview of all loan applications and their status.
- 2.3. As a bank employee, I want the system to generate reports regarding approval rates, rejection rates, and average processing times, so that I can assess how efficiently the loan applications are processed.
- Scenario 1: Model Performance and Retraining
- As a developer, I want to submit new data batches for model assessment and retraining to ensure the AI's performance remains optimal.
- Scenario 2: Monitoring and Updating the Model
- As a developer, I want to monitor the model's performance metrics and update the model to newer versions as needed.
- Akuen Akoi Deng
- Andreea Lavinia Fulger
- Cynthia Tarwireyi
- Daniel Dovhun
- Kanokwan Haesatith
- Nazli Moghaddam
- Djangos
- Pandas
- Tensorflow
- scikit-learn
- Dockers and Kubernetes
- SQLlite
- Trello Board
For more documentation visit the project wiki