Caravan Insurance Customer Identification

This project focuses on classifying potential customers for buying Caravan insurance. The target variable is Caravan (Binary Classification).

Files

• Dataset:

  • caravan.csv

• Python Code:

  • Jupyter notebook: caravan.ipynb

Problem Statement

The goal is to classify potential customers into two categories based on their likelihood to purchase Caravan insurance. The target variable is binary (0 or 1).

Kaggle Dataset

The dataset contains 87 variables and 9823 observations. Refer to the Kaggle link for variable descriptions: Caravan Insurance Customer Identification.

Data Analysis

• Data Splitting:

  • The dataset was divided into two data frames, Train and Test, using the ORIGIN variable. The variable was then dropped.
  • Train Shape: (5822, 86)
  • Test Shape: (4000, 86)

• Down Sampling:

  • Due to imbalanced classification, down sampling was applied.
  • Original Value Counts: 0 - 3762, 1 - 238
  • Downsampled Value Counts: 0 - 400, 1 - 348

• No Missing Values

• Feature Selection:

  • Applied chi-sq-Test for Categorical variables with a p-value threshold limit of 0.02.
  • Reduced dimensions from 87 to 26.

• Model Development:

  • Initially, the model achieved 75% accuracy.
  • After reducing the p-value of the chi-sq test to 0.02, the model achieved 100% accuracy with no False Positives (FP) and False Negatives (FN).

  Accuracy: 1.0 Confusion Matrix:

  [[3762 0] 
  [0 238]]
  

Usage

1. Clone the repository:

   git clone <repository_url>

2. Navigate to the project directory:

   cd Caravan-Insurance-Customer-Identification

3. Explore the dataset and Jupyter notebook for detailed analysis and code.

Feel free to refer to the caravan.ipynb for an in-depth explanation of the project.

---

Note: Adjust file paths and comments as needed for your project structure.