A Meta-classifier Approach for Medical Diagnosis
| Autor: | Andreas Stafylopatis, Konstantina S. Nikita, Dimitrios S. Frossyniotis, George L. Tsirogiannis |
|---|---|
| Rok vydání: | 2004 |
| Předmět: |
Boosting (machine learning)
Artificial neural network business.industry Computer science Decision tree Machine learning computer.software_genre Support vector machine Random subspace method ComputingMethodologies_PATTERNRECOGNITION Artificial intelligence Medical diagnosis business Classifier (UML) computer |
| Zdroj: | Methods and Applications of Artificial Intelligence ISBN: 9783540219378 SETN |
| DOI: | 10.1007/978-3-540-24674-9_17 |
| Popis: | Single classifiers, such as Neural Networks, Support Vector Machines, Decision Trees and other, can be used to perform classification of data for relatively simple problems. For more complex problems, combinations of simple classifiers can significantly improve performance. There are several combination methods, like Bagging and Boosting that combine simple classifiers. We propose, here, a new meta-classifier approach which combines several different combination methods, in analogy to the combination of simple classifiers. The meta-classifier approach is employed in the implementation of a medical diagnosis system and evaluated using three benchmark diagnosis problems as well as a problem concerning the classification of hepatic lesions from computed tomography (CT) images. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|