Hand written digit recognition using Logistic Regression, kernel SVM with PCA/LDA dimensionality reduction, and Deep Neural Network (Lenet-5 architecture) for MINST dataset.

python3 main.py --Method Lenet

• Parameters
  • Method - Classifiers. Default :- 'Lenet'

python3 main.py --Method LR --DimRed LDA 

• Parameters
  • Method - Classifiers. Default :- 'Lenet'
  • DimRed - Dimensionality Reduction technique. Option : 'PCA/LDA' Default :- 'PCA'

python3 main.py --Method SVM --DimRed LDA --Kernel Polynomial

• Parameters
  • Method - Classifiers. Default :- 'Lenet'
  • DimRed - Dimensionality Reduction technique. Option : 'PCA/LDA' Default :- 'PCA'
  • DimRed - kernel for Kernel SVM. Option : 'Polynomial/RBF' Default :- 'Linear'

Accuracy	Confusion Matrix

Dim Red.	Accuracy	Confusion
PCA
LDA

Kernel	PCA	LDA
Linear
Polynomial
RBF