Each year, banks are inundated with hundreds of credit card applications. The approval process for each application relies on several crucial factors, such as income level, delinquency records, and credit history. However, manually reviewing every single application can be an tedious, error-prone, and time-consuming task. Fortunately, the power of machine learning offers an automated solution to streamline this process. Nowadays, the majority, if not all, banks utilize machine learning algorithms to detect credit card approval. In this project, I aim to develop a credit card approval prediction by using machine learning techniques.
- The data
- Jupyter Notebook
- Load the data
- Take a quick look at the data structure
- Splitting the dataset into train and test sets
- Stratify sampling
- Discover and Visualize the Data to Gain Insights
- Exploratory data analysis for categorical attributes
- Separating the categorical and numerical attributes
- Frequency Distribution
- Target Variable Analysis
- Analyzing the correlation between categorical attributes and the target variable
- Exploratory data analysis for numerical attributes
- Summarizing numerical attributes
- Analyzing the correlation between numerical attributes and the target variable
- Visualizing the correlations between numerical attributes
- Exploratory data analysis for categorical attributes
- Handling missing values
- Missing values for categorical attributes
- Missing values for numerical attributes
- Preprocessing the data
- Feature selection
- Highly correlated features
- Important features
- Preparing the data for training
- Trying out different models
- Models predictions and evaluation
- Fine-tuning the most promising models
- Evaluating the best performing model on the test set
- Conclusion