Machine Learning on Classic Dataset

1. MNIST 
• Logistic Regression: 7.63% Error - Very fast - pylearn2
• Random Forest: 3% Error - Very fast (sklearn) 
• MLP - 2% Error
2. CIFAR10 
• Logistic Regression: 17% Accuracy 
• PCA + Logistic Regression: 39% Accuracy 
• Adaboost: 40% Accuracy  
• Random Forest: 49% Accuracy 
• GBM: 50% Accuracy (very slow)