RADAR

Sorry it's just the bare minimum so far.
Clare, A., Williams, H. E. and Lester, N. M. (2004) Scalable Multi-Relational Association Mining. In proceedings of the 4th International Conference on Data Mining ICDM '04. See pdf

Radar is based on the technology from RMIT's Zettair search engine.

Use

(-h for help)

usage: ./radar [-a] [-p] [-i] [-m bytes] [-f index_name]
       [-c [-d max_depth] [-s min_support] [-r] [-n startfrom] [-t]

        -i: use ./radar to produce an index
        -a: auto mine
        -m: limit memory usage to approximately this amount (in bytes)
        -f: use this as the prefix for the index name
        -p: prints the index
        -c: count all associations to max_depth, using min_support frequency, 
optionally restarting (-r) from an assoc given on stdin, or starting counting 
from a particular term in bias
        -t: tests an indexed database against an assoc file, read from stdin

Source

C source for Linux: radar.20040715.tgz

Install with:

tar xvfz radar.20040715.tgz
cd radar.20040715
make

Sample data

Mutagenesis: muta.tgz