This project aims to develop a machine learning model that predicts whether an applicant should be approved for a credit card based on various socio-economic and financial factors. The goal is to use different classification algorithms to optimize accuracy and provide insights into key approval criteria.
The dataset contains information about credit card applicants, including their demographics, financial status, and employment details. The features considered for prediction are:
Features:
- Applicant_Gender: Gender of the applicant
- Owned_Car: Whether the applicant owns a car
- Owned_Realty: Whether the applicant owns real estate property
- Total_Children: Number of children the applicant has
- Total_Income: Applicant's annual income
- Income_Type: Type of income source
- Education_Type: Applicant's highest education level
- Family_Status: Marital status of the applicant
- Housing_Type: Applicant's living arrangements
- Owned_Mobile_Phone: Whether the applicant owns a mobile phone
- Owned_Work_Phone: Whether the applicant owns a work phone
- Owned_Phone: Whether the applicant owns a landline phone
- Owned_Email: Whether the applicant owns an email address
- Job_Title: Applicant’s current job title
- Total_Family_Members: Number of family members in the applicant's household
- Applicant_Age: Applicant’s age
- Years_of_Working: Total years of employment
- Total_Bad_Debt: Number of previous bad debts
- Total_Good_Debt: Number of previous successfully repaid debts
Data Cleaning:
- Handling missing values by either removing records with significant missing data or imputing values.
- Converting categorical variables into numerical format using one-hot encoding.
- Standardizing numerical features such as income, age, and years of working .
Multiple machine learning algorithms are tested to find the best-performing model
Algorithms Used:
- Logistic Regression
- Decision Tree Classifier
- Random Forest Classifier
- XGBoost