đŸ”ĸ MNIST Digit Recognition with Naive Bayes

Draw a digit (0-9) in the canvas below and the model will predict what digit it is!

Model Details:

  • Dataset: MNIST (Keras/TensorFlow - 28x28 pixels)
  • Algorithm: Multinomial Naive Bayes
  • Accuracy: 83.65%

📊 How to use:

  1. Draw a digit (0-9) in the canvas on the left
  2. Click "Predict Digit" button
  3. See the prediction results. If it's not a clear digit, it may show "Unknown".

Model Performance Metrics:

  • Training Samples: 60000
  • Testing Samples: 10000
  • Model Accuracy: 83.65%
  • Algorithm: Multinomial Naive Bayes
  • Confidence Threshold: 99%