# Apply linear Discriminant analysis

Let's say that our feature space is of dimension size = 9. Also, let's say that we apply LDA and we get only one LDA component. If we developed for example the SVM model and test the accuracy with these 9 features and then do the same with this one LDA component and calculate this new accuracy. IF the accuracy with the 9 features = the accuracy with this one LDA component, then this means that all the 9 features are important ?