- This project focuses on predicting credit card defaults using machine learning techniques. We’ll walk through the steps from data preparation to model deployment.
- Data Acquisition
Obtained the dataset from the UCI Machine Learning Repository.
Explored the metadata to understand its structure and features. - Data Cleaning
Cleaned the dataset by handling missing values, duplicates, and outliers.
Ensured data consistency and integrity. - Exploratory Data Analysis (EDA)
Conducted EDA to gain insights into the data.
Visualized distributions, correlations, and patterns. - Preprocessing
Imputed missing values using appropriate techniques (mean, median, etc.).
Checked for duplicate records.
Prepared the data for model training. - Model Selection
Split the dataset into training and testing sets.
Normalized features to ensure consistent scaling.
Experimented with various machine learning models:- Logistic Regression
- Decision Tree Classifier
- Random Forest Classifier
- Model Evaluation
Evaluated model performance using metrics such as accuracy, precision, recall, and F1-score.
Selected the best-performing model based on validation results. - Model Deployment
Saved the best model as a .pkl file for future use.
Created a .py file for the user interface using Streamlit to allow real-time predictions based on user input.
Clone this repository and install the required dependencies.
Load the pre-trained model using the .pkl file.
Use the .py file to make a webpage to predict credit card defaults based on real-time user input.
Feel free to customize this template with specific details about your dataset, features, and findings.
Good luck with your project! 🚀