Naive Bayes classifier for Iris dataset with custom probability estimation. Built using pandas and numpy.
This project implements a Naive Bayes classifier for the Iris flower dataset. The probability estimation is custom-built without using scikit-learn or other ML libraries.
- Custom probability estimation function
- Data preprocessing and splitting (70% train, 30% test)
- Bayesian classification implementation
- Model evaluation metrics
- python
- pandas==2.0.0
- numpy==1.21.0
- matplotlib==3.5.0
Training accuracy: ~78-92% Test accuracy: ~75-90%