classification - Is my Matlab code correct for applying PCA to data? -


I have the following code for calculating PCA in Matlab: 

  train_out = Train '; Test_out = test '; Decrease the mean off for% each dimension mn = mean (train_out, 2); Train_out = train_out - repmat (mn, 1, train_size); Test_out = test_out - repmat (mn, 1, test_size); Calculation of% sympathetic matrix co-operation = 1 / (train_size -1) * train_out * train_out '; % Eigenvectors and eigenvalues ​​get [pc, v] = eig (co-type); Removing the diagonal of the matrix as Vector V = Dye (V); Variations in order decreasing [junk, rindices] = sort (-1 * v); V = rindices; PC = PC (:, rindices); % Project Basic Data Set = PC '* train_out; Train_out = out '; Out = PC '* test_out; Test_out = out ';

The train and test matrix lines have a feature variable in observation and columns. When I classify the original data (without PCA), I get better results than PCA, even when I have all the dimensions. When I tried to straighten the PCA on the whole dataset (train + test) then I saw that the interrelationship between these new principal components and the previous ones are either about 1 or about -1 which makes me feel weird. I'm probably doing something wrong, but I can not understand it.

The code is correct, although using the precomcump function can be easy:

< Pre> train_out = train; To save% original data, test_out = test; MN = means (train_out); Train_out = bsxfun (@ zero; train_out, mn); % Sub abstract means test_out = bsxfun (@ zero, test_out, mn); [Coffs, scores, variance] = prinxp (train_out, 'econ'); % PCA pervar = cumsum / variances; Dims = Max (search (pervar & lt; var_frac)); % Var_frac - e.g. 0.99 - explained the fraction of the variance Train_out = Train_out * Coff ((, 1: dim);% Dims - Keep this multiple dimensions test_out = test_out * coefs (:, 1: dims);% result train_out and test_out

-

Comments

Post a Comment

Popular posts from this blog

Member with no value in F# -

After I am trying to convert F # to the following C # code: [StructLayout (LayoutKind.Sequential, charset = CharSet.Ansi)] Private class DOCINFOA {[MarshalAs (UnmanagedType.LPStr)] public string nomeDocumento; [Marshall (UnmanagedType.LPStr)] Public string arquivoSaida; [Marshall (UnmanagedType.LPStr)] Public String Tipodo; } I am doing the following: Named place printer module RawPrinterHelper = [& LT; StructLayout (LayoutKind.Sequential, charset = CharSet.Ansi) & gt;] Type DOCINFOA = [& LT; MarshalAs (UnmanagedType.LPStr) & gt;] Member this.nomeDocumento [& LT; MarshalAs (UnmanagedType.LPStr) & gt;] Member this.arquivoSaida [& LT; MarshalAs (UnmanagedType.LPStr) & gt;] Member This Tipoda [/ code> But I get: One type of definition is necessary for one or more members . Can I vacate these members? you val code> member : [& lt; StructLayout (LayoutKind.Sequential, charset = CharSet.Ansi) & gt;] Type DOCINFOA = [...
Read more

java - Joda Time Interval Not returning what I expect -

मेरे पास निम्न प्रोग्राम है import java.util। *; Import java.text। *; आयात करें org.joda.time। *; सार्वजनिक श्रेणी के स्कोप कंट्रोल {सार्वजनिक स्थिर अंतराल मिलनसार () {दिनांक समय currDate = नया दिनांक समय (2008, 4, 4, 15, 30, 0, 0); दिनांक समय epochDate = नया दिनांकटाइम (2000, 1, 1, 12, 0, 0, 0); अंतराल अंतराल = नया अंतराल (युरोप डेट, करोडेट); वापसी अंतराल; } सार्वजनिक स्थिर शून्य मुख्य (स्ट्रिंग [] आर्ग्स) {डबल दिनबात = मिलते समय ()। ToDurationMillis () / 1000/60/60/24; StdOut.println (daysBtween); }} मुझे आउटपुट मिल रहा है: 3016.0 लेकिन जो मैं देख रहा हूं वह है: 3016.1458333333 मैं क्या कर रहा हूँ ? toDurationMillis एक लंबा लौटा देता है, और प्रत्येक प्रभाग को int के रूप में घोषित किया जाता है। जावा इस प्रकार इनट्स को लॉन्ग में परिवर्तित कर देगा और डिवीजन को लंबे समय तक लौटाना होगा। परिणाम को अंत में दोहरे रूप में परिवर्तित करना। डबल्स का उपयोग करके अभिव्यक्ति करने के लिए जावा को बताने के लिए, अभिव्यक्ति के किसी भी घटक को दोहरे रूप में घोषित करें। उदाहरण के लिए: ...
Read more

c# - Showing a SelectedItem's Property -

So I have enabled a rightclick option for my DataGrid. I want to display the only chosen property but it is not behaving like what I would like it to display my name and additional. Public category Paymentinfo {public int PaymentNo {get; Set; } Public string date {get; Set; } Public Double Pay {get; Set; } Public Double Principle {get; Set; } Public Double Interest {get; Set; } Public Double Balance {Received; Set; }} Private Zero MenuItem_OnClick (Object Sender, RoutedEventArgs e) {MessageBox.Show (AmortGrid.SelectedItem.ToString ()); } I'm trying to apply it without using a View Module! If I place a breakpoint where the messagebox and the recipient place it on selected ITEM, then this property will display the payment. A date-payment-principle-interest-balance is the only value that needs payment. Was hoping that it would be something like this MessageBox.Show (AmortGrid.SelectedItem.PaymentNo.ToString ()); When you call ToString () So you get the class typ...
Read more