MNIST Digit Classification using K-Nearest Neighbors (KNN)

This is an educational project designed to demonstrate the implementation of the K-Nearest Neighbors (KNN) algorithm on the MNIST dataset. It is a simple, beginner-friendly project aimed at learning the basics of machine learning and classification tasks. I learned the concepts of machine learning by working on this project and implemented it myself. It focuses on classification using the K-Nearest Neighbors (KNN) algorithm, and through this, I gained hands-on experience with key tasks such as data loading, preprocessing, model training, and evaluation.

Key Steps:

1. Loading the Dataset: The MNIST dataset, which contains images of handwritten digits, is loaded using fetch_openml.

2. Data Preprocessing:

   • The data is split into training and testing sets.
   • Standardization is applied to scale the pixel values using StandardScaler for better model performance.

3. Model Training:

   • The KNN classifier is trained on the preprocessed data to classify the digits.

4. Model Evaluation:

   • The model is evaluated using cross-validation, and its performance is measured with accuracy and a confusion matrix.

5. Saving the Model:

   • After training, the model is saved using joblib, so it can be reused later without retraining.

Requirements:

To run the code, you need to install the following libraries:

pip install scikit-learn numpy matplotlib seaborn joblib