Rule generation was done using the same options for RGEN as above.
As mentioned, only RGEN and RCLASS was used to generate test
results for these data. Table 
shows which
classification criteria were used.
| 1 | RGEN - Simple voting, use only top node in lattice. Close to RSES |
| 2 | RGEN - Highest accuracy, all nodes |
| 3 | RGEN - Simple voting with linear weight, all nodes |
| 4 | RGEN - Simple voting with exponential weight correction of accuracy |
| 5 | RGEN - Simple voting with squared weight correction |
| 6 | RGEN - Measurement of evidence method |
| 7 | Simple voting, linear weight, 2 upper levels of nodes |
| 8 | Simple voting, linear weight, 2 lower levels of nodes |
Four different ``families'' of data sets were used for the experiment, with characteristics as shown below.
| Data set | rel. attr. | irrel. attr. | dom.size | classes | max | noise | train. | test | rule conjs. |
| Small noise-free | 4 | 0 | 4 | 4 | 50% | 0% | 30 | 70 | 1-3 |
| Small, 10% noise | 4 | 0 | 5 | 10 | 25% | 10% | 30 | 70 | 0-2 |
| Med., 20% noise | 6 | 0 | 3 | 15 | 15% | 20% | 50 | 250 | 1-3 |
| Large, 5% noise | 7 | 1 | 5 | 10 | 25% | 5% | 75 | 225 | 1-3 |
shows classification results on a small data set
in which no noise is added.
In Table , 10% of the values among the condition
attributes are given a random value. 20% noise was added to the
medium sized data set tested in Table , while only 5% in the
large data set with test results in Table .
| Set | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| 1 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 97.1 |
| 2 | 100.0 | 98.5 | 98.5 | 98.5 | 98.5 | 98.5 | 100.0 | 75.7 |
| 3 | 100.0 | 97.1 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 88.6 |
| 4 | 100.0 | 97.1 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 84.3 |
| 5 | 100.0 | 91.4 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 91.4 |
| Set | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| 1 | 71.4 | 50.0 | 72.9 | 72.9 | 74.3 | 72.9 | 72.9 | 57.1 |
| 2 | 65.7 | 50.0 | 65.7 | 65.7 | 64.3 | 65.7 | 64.3 | 52.9 |
| 3 | 71.4 | 62.8 | 65.7 | 65.7 | 67.1 | 77.1 | 68.6 | 52.9 |
| 4 | 77.1 | 68.6 | 78.5 | 78.5 | 78.5 | 78.5 | 77.1 | 72.8 |
| 5 | 70.0 | 65.7 | 70.0 | 71.4 | 68.5 | 74.3 | 70.0 | 72.9 |
| Set | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | |
| 1 | 51.6 | 39.6 | 50.0 | 50.0 | 48.8 | 50.4 | 48.8 | 28.0 | |
| 2 | 54.8 | 40.4 | 53.2 | 52.4 | 54.4 | 51.2 | 52.8 | 22.8 | |
| 3 | 56.0 | 33.2 | 58.6 | 60.0 | 58.8 | 61.2 | 57.2 | 38.8 | |
| 4 | 62.8 | 38.8 | 62.0 | 61.2 | 62.8 | 60.8 | 61.6 | 38.8 | |
| 5 | 63.2 | 40.4 | 60.4 | 60.4 | 59.2 | 62.0 | 62.8 | 28.0 |
| Set | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| 1 | 78.7 | 20.4 | 78.7 | 78.7 | 77.8 | 77.3 | 78.7 | 65.8 |
| 2 | 74.7 | 31.1 | 76.9 | 76.0 | 78.7 | 77.8 | 74.7 | 56.4 |
| 3 | 79.0 | 29.5 | 78.1 | 78.6 | 78.6 | 74.1 | 79.0 | 65.2 |
| 4 | 79.1 | 29.7 | 78.2 | 77.8 | 78.7 | 74.2 | 79.1 | 64.9 |
| 5 | 68.9 | 24.4 | 64.9 | 64.9 | 62.2 | 66.2 | 68.9 | 56.4 |