Hey Dr. Menzies, Here there is some raw data from the tests I am performing. I do not have graphs yet because the tests have not finished, but I will have them for tomorrow. The files included are (within a zip file): teston10n.txt: Comparison between NBC with gaussian, NBK and NBC with SPADE. teston10c.txt: Comparison between C4.5 and C4.5 with SPADE prediscretized datasets. 10by10.wlt: win-loss table of independent datasets. recurrent.txt: a SAWTOOTH run over the simulation data repeated 3 times. kddBinNumberAbc.txt: A run of ABC using SPADE on the KDD cup 99 dataset of approximately 5 million records for training and 311029 instances for testing. It shows the number of bins created by SPADE. phyBinNumberSt.txt : Sawtooth on the physics dataset from the KDD cup 2004. It has 50 thousand records. Only training was performed. A list and number of bins was collected in this test. I hope the tests finish tonight and I will send you graphs tomorrow. Andres P.S. Results of SPADE on NBC look good!