This program generates a visualization of the clustering algorithm presented in Jason Baldridge LIN 313 class, SPRING 2013 at the University of Texas at Austin.
It's easy to run and all packages are included. You just need to have python 2.* running on your computer.
make
Runs a quick demo. You can just type this on your terminal (once you've navigated to the directory with these files).
By default, the program will load 100 randomly placed documents and 2 randomly placed centroids (categories).
make demo DOCS=200 CATS=3
You can specify how many documents and categories (or centroids) to visualize. For example, this will test for 200 documents and 3 categories.
make file FILE=data.txt
You can also specify a file with data for visualization. The file structure is as follows:
4
123,234
345,123
...
The first line determines the number of categories, or centroids. after that, the program will look for coordinates points with domain and range of [0,500] separated with a comma and no spaces.
We provide 4 example files: data1.txt, data2.txt, data3.txt, and data4.txt.
make demo
make demo DOCS=400 CATS=7
make file FILE=data.txt
make random
These commands will specifically call for the cluster python program.
make random will create a randomly generated data.txt file.
make clean
This command will delete unnecessary files and clean your directory.
Now you can visualize the clustering process with Python!