Question

I have a slight problem, I cant reverse the singular value decompisition from my process but I was wondering if I can map data beforehand.

One of my datasets is named fulldata. I apply svds to this data like so:

%% dimensionality reduction 
columns = 6
[U,S,V]=svds(fulldata,columns);

I then randomly pick 1000 rows from the dataset:

rows = 1000;
columns = 6;

%# pick random rows
indX = randperm( size(fulldata,1) );
indX = indX(1:rows);

%# pick random columns
indY = indY(1:columns);

%# filter data
data = U(indX,indY);

I need to find a way in which I can tell which 1000 rows it picked from the fulldata? Maybe output data from 1 – 1000 with the row number from fulldata. Does anyone know a way in which it can be done?

Answer 1

Actually you are almost there:

rows = 1000;
indX = randperm( size(fulldata,1) );
indX = indX(1:rows);

dataSample = fulldata(indX, :); 

dataSample will now contain all rows of fulldata specified in indX.