This project classifies Near-Earth Objects (NEO) and Potentially Hazardous Asteroids (PHA) using a Naive Bayes classifier on a PostgreSQL database. Data cleaning and feature engineering are done in PostgreSQL, with Liquibase for database versioning, and the classifier is implemented manually in R.

dataset: kaggle

The project uses the Naive Bayes classifier to identify and categorize asteroids based on orbital and physical parameters. Key steps include:

• Data Cleaning: Processed in PostgreSQL for consistency and completeness.
• Feature Engineering: Binning key features to simplify classification.
• Database Management: Using Liquibase to track database changes.
• Manual Naive Bayes in R: Implemented in R to classify asteroids as NEO or PHA.

Asteroid data features include:

• Moid (Minimum Orbit Intersection Distance)
• a (Semi-major axis)
• e (Eccentricity)
• i (Inclination)
• H (Absolute magnitude)

These features are grouped into bins to create categorical data for classification.

• PostgreSQL: For data storage and cleaning.
• Liquibase: For database migrations.
• R and RStudio: For implementing the Naive Bayes classifier.

1. Clone the repository and navigate to the project directory.
2. Database Setup: Create a PostgreSQL database, then use provided changelogs scripts to set up tables and load data.
3. Liquibase Migrations: Run migrations using Liquibase for version control.
4. Run Classifier in R: Execute the Naive Bayes classification in R to predict NEO and PHA categories.

Classification results and additional analyses are saved and visualized in R, providing insights into asteroid characteristics.