Home Credit is currently using various statistical methods and Machine Learning techniques to make credit score predictions. Now, as a Data Scientist Intern, we will unlock the full potential of our data. By doing so, we can ensure that customers who are able to repay are not rejected when applying for loans, and that loans are provided with a principal, maturity, and repayment calendar that will motivate customers to succeed.
Home Credit Indonesia is facing several challenges:
- Ensuring that customers who can repay their loans (class 0) are not rejected to increase profitability.
- Identifying customers who are likely to face repayment issues to minimize loans granted to unqualified customers and reduce potential losses.
This dataset 307,511 customers who have taken loans, with 122 features such as ID, gender, income, etc.
To address these challenges, a machine learning model will be developed to predict credit scores, helping determine clients capable of repaying loans and those who may encounter payment issues.
- Identify the main problems to be solved.
- Select appropriate machine learning models and evaluation metrics.
- Conduct exploratory data analysis (EDA) to understand the dataset characteristics and uncover valuable insights.
- Perform data preprocessing, including data cleansing, feature encoding, handling class imbalance, etc.
- Build classification models, such as logistic regression, decision tree, or others.
- Determine the best model based on evaluation metric results.
- Provide business recommendations based on analysis and findings.
- Balanced Accuracy: The average recall for class 0 and class 1, ensuring good performance for both classes.
- Recall: Measures how well the model identifies all positive cases out of the actual positives in the data.
- Precision: Measures the accuracy of the model in predicting the positive class, indicating how many predicted positives are truly positive.
- Python Programming Language
- JupyterLab / Jupyter Notebook