Continuous Attribute Discretization Based on Inflection Point
| Autor: | Kewen Xia, Liu Liu, Jianchuan Bai, Yue Chi |
|---|---|
| Rok vydání: | 2014 |
| Předmět: |
Mathematical optimization
Discretization Mathematical definition Discretization algorithm Library and Information Sciences Discretization error Computer Graphics and Computer-Aided Design k-nearest neighbors algorithm Computational Theory and Mathematics Inflection point Compatibility (mechanics) Applied mathematics Information Systems Discretization of continuous features Mathematics |
| Zdroj: | Journal of Information and Computational Science. 11:1327-1333 |
| ISSN: | 1548-7741 |
| DOI: | 10.12733/jics20103079 |
| Popis: | Continuous attribute discretization plays an important role in data mining. A novel discretization method based on inflection point of curve is presented, which includes calculating the all inflexion points as a candidate discrete point based on the mathematical definition of inflection point, then using decision attribute types of sample as constraint condition, and obtaining the final discrete segment points by Nearest Neighbor Rule. After the actual processing for logging attributes discretization and comparative analysis, the results show that the compatibility of presented discretization method is high in the condition of keeping indiscernibility relation, and its effect is superior to that of the discretization algorithm based on attribute importance. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|