TreeGen Algorithm

Contains the synthetic tree generator described in [2005-treeminer:tkde]

Relevant Publications

• [2005-treeminer:tkde] Mohammed J. Zaki. Efficiently mining frequent trees in a forest: algorithms and applications. IEEE Transactions on Knowledge and Data Engineering, 17(8):1021–1035, August 2005. special issue on Mining Biological Data. doi:10.1109/TKDE.2005.125.

HOW TO

1. type tree_gen to see the options.

    -d : Depth, default 5
    -f : Fan out factor, default 5
    -a : ascii output flag
    -b : use random subtree root as starting root
    -n : Number of items, default 10
    -m : Total number of nodes in parent tree
    -o : out file name (required)
    -p : output parent tree as database
    -s : seed for random number gen
    -t : Number of subtrees, default 100      
   

The program first generates a master tree with total m nodes with avg depth d and fanout f from n items (or labels)

2. you can output this parent tree in the file name specified by -o (and with -a if you want ascii output), by specifying -p option

3. you can generate t subtrees from the parent by omitting -p option and specifying the file name -o (with -a for ascii).

example run:

tree_gen -f 6 -d 10 -n 100 -m 1000 -o xxx -a -t 1000

to produce a pool of 1000 trees from the master.

If you want to know the master tree, run:

tree_gen -f 6 -d 10 -n 100 -m 1000 -o xxx.master -a -t 1000 -p