Handwritten-Digit-Recognition-with-Logistic-Regression

Trains and evaluates a logistic regression model for handwritten digit recognition using both scikit-learn’s load_digits (8×8) and MNIST (28×28) datasets. Includes manual parameter tuning, normalization, and dataset comparison. Achieved 99.2% accuracy on load_digits and 92.2% on MNIST, with detailed evaluation via accuracy, F1 score, confusion matrix, and visualizations.
https://www.twine.net/signin