Table 6.
The five datasets used to evaluate the rules parsed from the InterPro database.
| the parsed rules# | |||
| precision | Confidence | coverage* | |
| A testing data | 52% | 48% | 23% |
| B testing data | 51% | 50% | 24% |
| C testing data | 56% | 52% | 26% |
| D testing data | 47% | 41% | 20% |
| E testing data | 64% | 62% | 47% |
A: actinobacteria B: bacillales C: fungi D: nematode + arthropoda E: viridiplantae
*: coverage = the hit ratio of testing data
#: The dataset was parsed from the entry xref table of the InterPro database. The IPR Acc's were corresponding to ENZYME.