Example 2 to visualize topic model results

This example shows how to visualize documents in a 2D map. The documents are located such that neighboring documents have similar topic distributions. Documents are shown by their corresponding titles

Load a document-topic count matrix saved for the nips dataset

load 'ldasingle_nips';

Get the title descriptions of the papers

load 'titles_nips';

We cannot visualize every paper, take a random subset

SEED = 5;
rand( 'state' , SEED ); subs = randperm( size( DP,1));

NS = 200;

These are the indices of the random subset of documents

subs = subs( 1:200 );

Visualize the documents of this subset

NCHARS   = 30; % maximum number of characters on one line
MAXLINES = 3; % maximum number of lines

fprintf( 'Please wait while calculating visualization...\n' );
drawnow;

VisualizeDocs( DP(subs,:) , ALPHA , titles(subs) , NCHARS , MAXLINES )

Please wait while calculating visualization...