| 1. |
|
|
| 2. |
|
|
| 3. |
- Lindgren, Tony, et al.
(author)
-
Classification with Intersecting Rules
- 2002
-
In: Proceedings of 13th International Conference on Algorithmic Learning Theory (ALT'02). - Lübeck : Springer Verlag. ; , s. 395-402
-
Book chapter (other academic/artistic)
|
|
| 4. |
- Lindgren, Tony
(author)
-
Methods for rule conflict resolution
- 2004
-
In: MACHINE LEARNING. - BERLIN : SPRINGER. - 3540231056 ; , s. 262-273
-
Conference paper (peer-reviewed)abstract
- When using unordered rule sets, conflicts can arise between the rules, i.e., two or more rules cover the same example but predict different classes. This paper gives a survey of methods used to solve this type of conflict and introduces a novel method called Recursive Induction. In total nine methods for resolving rule conflicts are scrutinised. The methods are explained in detail, compared and evaluated empirically on an number of domains. The results show that Recursive Induction outperforms all previously used methods.
|
|
| 5. |
|
|
| 6. |
- Lindgren, Tony, et al.
(author)
-
Resolving rule conflicts with double induction
- 2004
-
In: Intelligent Data Analysis. - : IOS Press. - 1088-467X .- 1571-4128. ; 8:5, s. 457-468
-
Journal article (peer-reviewed)abstract
- When applying an unordered set of classification rules, the rules may assign more than one class to a particular example. Previous methods of resolving such conflicts between rules include using the most frequent class of the examples covered by the conflicting rules (as done in CN2) and using naïve Bayes to calculate the most probable class. An alternative way of solving this problem is presented in this paper: by generating new rules from the examples covered by the conflicting rules. These newly induced rules are then used for classification. Experiments on a number of domains show that this method significantly outperforms both the CN2 approach and naïve Bayes.
|
|
