Implement PCA and LDA from scratch. Classify the dataset as it is and use the same classification algorithm for the parts below. Use Linear discriminant analysis (LDA) to find out the best projection directions. Project your data on the new projection matrix given by LDA. Classify the projected data using 5 fold cross-validation, report the mean and standard deviation of classification accuracy. Use the best model, to classify the test set and plot the ROC curve and confusion matrix. Perform PCA on the same data by preserving 95% eigen energy and report the mean accuracy and standard deviation over the 5 folds.Compare and analyze the results obtained by PCA and LDA. Visualise and analyse the eigenvectors obtained using PCA If you perform LDA on the PCA projected data, find out the classification performance for both the databases.

1. Dataset 1 : https://drive.google.com/file/d/10p0t4Kbrz--UIAn_nfys-T5HeHjM8hn8/view
2. Dataset 2 : https://www.cs.toronto.edu/~kriz/cifar.html