Mnist_project

---

Working with Mnist dataset usiing different models and machine learning techniques.

---

Train a binary classifier in order to distinguish between two different classes

• k-fold cross alidation
• Confusion matrrix
• ROC curve, AUC

Use PCA and implementation of Ensemble classifier

• PCA
• Decision Tree, Random Forest, Adam Boost, Linear SVC, Logistic Rregression
• Stacking Ensemble Classifier and Random Forest Classifier
• k-fold cross validation
• Score of Stacking Classifier

Keep specific digits using PCA and clustering

• PCA and visualization of two components
• K-Means clstering

Build a feed forward neural network using keras

• Adam optimizer
• Plot history of loss and accuracy of model
• Regularization model to avoid oerfitting

Build convolutional neural network (CNN)

• 2D conolutional layers
• 2x2 max ppooling layer
• Adam optimizer
• Monitor validation set
• Plot history of loss and accuracy

Build a vanilla auto-encoder

• Monitor validation set
• Calculation of reconstruction error