Brain Tumor Classification with Naive Bayes

This exploratory machine learning project investigates the classification of brain tumors using features extracted from MRI images. It focuses not only on building a predictive model, but also on evaluating how confident the model is in its decisions, a key aspect in sensitive domains like medicine.

---

Dataset

The dataset contains:

• 700 samples
• 1500 numerical features (extracted from MRI scans)
• 4 tumor types:
  1. Pituitary Adenoma
  2. Germinoma
  3. Meningiomas
  4. Glioma

Source: Kaggle – Brain Cancer Data

---

Approach

The project includes:

• Data exploration & visualization
• Feature selection using ANOVA F-statistic (SelectKBest)
• Naive Bayes classification (GaussianNB)
• Cross-validation to test generalizability
• Prediction probability analysis
• Probability calibration with CalibratedClassifierCV
• Visual inspection of model uncertainty

---

Highlights

Section	Description
Feature Exploration	KDE plots to visualize distribution overlaps and redundancy
Feature Selection	Top 500 most informative features selected
Model	Gaussian Naive Bayes classifier
Cross-Validation	5-fold CV for robustness
Calibration	Platt scaling (sigmoid) to correct overconfidence
Visuals	Confusion matrix, prediction bars, confidence histograms

---

Key Insight

Without calibration, Naive Bayes often predicted probabilities close to 1.0 even on incorrect predictions.
Calibration helped make probabilities more realistic and trustworthy, especially important for risk-aware domains like healthcare.

---

How to Run

1. Clone this repository
2. Make sure you have the following dependencies:

pip install numpy pandas matplotlib seaborn scikit-learn