32 lines
966 B
Mathematica
32 lines
966 B
Mathematica
|
function [] = Problem2(training_file, test_file)
|
||
|
|
|
||
|
|
% load data
|
||
|
|
training_data = load(training_file);
|
||
|
|
test_data = load(test_file);
|
||
|
|
|
||
|
|
% Remove label from data
|
||
|
|
training_labels = training_data(:, end);
|
||
|
|
training_data(:, end) = [];
|
||
|
|
test_labels = test_data(:, end);
|
||
|
|
test_data(:, end) = [];
|
||
|
|
|
||
|
|
% Visualize Eigenfaces
|
||
|
|
Eigenfaces(training_data, test_data);
|
||
|
|
|
||
|
|
% Visualize Eigenfaces
|
||
|
|
[neigenvectors] = ProportionOfVariance(training_data);
|
||
|
|
fprintf("Number of Eigenvectors that explain 90%% of the variance: %.8f\n",neigenvectors);
|
||
|
|
|
||
|
|
% plot kNN test errors using the number of eigenvector that explain 90%
|
||
|
|
% of the variance
|
||
|
|
ks = [1 3 5 7];
|
||
|
|
KNN_Error(neigenvectors, ks, training_data, test_data, training_labels, test_labels);
|
||
|
|
|
||
|
|
% perform back projection on first 5 images of training data using the
|
||
|
|
% first 10, 50, 100 PCA components
|
||
|
|
n_components = [10 50 100];
|
||
|
|
Back_Project(training_data, test_data, n_components);
|
||
|
|
|
||
|
|
end % end of function
|
||
|
|
|