The real-life data set for the purpose of running the experimental simulation was obtained from the relatively small database that contains the daily power consumption rates of some customers over a period of one year. the CIMEG project makes available such kind of data sets. approximately 5 Megabyte is the size of the database. For the experiment, only a subset
of the original data is used with 150000 tuples. The parameters used are:
c=75%,
minimum partition size epsilon=10
no. of partitions m=2048
allowable data change=+/-0.5%
Wateramrk used=16 bit watermark

to minimize the probability of decoding error, a special decodong technique was used. The watermark thus formed was subject to the normal attacks of deletion, alteration, and addition attacks. Multiple runs of the experiment was run and the results of these runs were averaged to obtain the conclusions. both uniform and normally distributed synthetic data were applied to the test and the results obtained.

For more details on the data sets used and the experiments, see:
http://citeseerx.ist.psu.edu/viewdoc/dow...1&type=pdf