Concept Compression in Formal Concept Analysis Using Entropy-Based Attribute Priority
Discovering important concepts in formal concept analysis (FCA) is an important issue due to huge number of concepts arising out of complicated contexts. To address this issue, this paper proposes a method for concept compression in FCA, involving many-valued decision context, based on information e...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2017-03-01
|
| Series: | Applied Artificial Intelligence |
| Online Access: | http://dx.doi.org/10.1080/08839514.2017.1316182 |
| Summary: | Discovering important concepts in formal concept analysis (FCA) is an important issue due to huge number of concepts arising out of complicated contexts. To address this issue, this paper proposes a method for concept compression in FCA, involving many-valued decision context, based on information entropy. The precedence order of attributes is obtained by using entropy theory developed by Shannon. The set of concepts is compressed using the precedence order thus determined. An algorithm namely Entropy based concept compression (ECC) is developed for this purpose. Further, similarity measures between the actual and compressed concepts are examined using the deviance analysis and percentage error calculation on the deviance of input weights of concepts. From the experiments, it is found that the compressed concepts inherit association rules to the maximum extent. |
|---|---|
| ISSN: | 0883-9514 1087-6545 |