Name
Abstract | ||
|---|---|---|
It is desirable to automatically learn the effects of actions in an unknown environment. C4.5 has been used to discover associations, and it can also be used to find causal rules. Its output consists of rules that predict the value of a decision attribute using some condition attributes. Integrating C4.5's results in other applications usually requires spending some effort in translating them into a suitable format. Since C4.5's rules are horn clauses and have the same expressive power as Prolog statements, we have modified standard C4.5 so it will optionally generate its rules in Prolog. We have made sure no information is lost in the conversion process. It is also possible for the prolog statements to optionally retain the certainty values that C4.5 computes for its rules. This is achieved by simulating the certainty values as the probability that the statement will fail for no apparent reason. Prolog can move from statement to statement and find a series of rules that have to be fired to get from a set of premises to a desired result. We briefly mention how, when dealing with temporal data, the Prolog statements can be used for recursive searches, thus making C4.5's output more useful. |
Year | DOI | Venue |
|---|---|---|
2000 | 10.1007/3-540-44491-2_13 | IDEAL |
Keywords | Field | DocType |
conversion process,apparent reason,speak prolog,logical decision rules,certainty value,prolog statement,standard c4,condition attribute,recursive search,expressive power,horn clause,causal rule,decision rule,temporal data | Decision rule,Programming language,Certainty,Horn clause,Computer science,Prolog,Temporal database,Artificial intelligence,Expressive power,Machine learning,Recursion | Conference |
ISBN | Citations | PageRank |
3-540-41450-9 | 3 | 0.47 |
References | Authors | |
2 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Kamran Karimi | 1 | 118 | 17.23 |
| Howard J. Hamilton | 2 | 1501 | 145.55 |