Optimization of clinical decision support based on Pearson correlation of attributes
Clinical decision support is very important especially in such a wide-spread disease as a coronary artery disease. A large variety of prediction methods can potentially solve the classification problem to support clinical decisions. However, not all of them provide similar efficiency for the classif...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Chapter/Article Conference Paper |
| Lenguaje: | inglés |
| Publicado: |
2019
|
| In: |
Phealth 2019
Year: 2019, Pages: 199-204 |
| DOI: | 10.3233/978-1-61499-975-1-199 |
| Acceso en línea: | Verlag: https://doi.org/10.3233/978-1-61499-975-1-199
|
| Notas de Autor: | Aleksei Dudchenko, Matthias Ganzinger, Georgy Kopanitsa |
| Sumario: | Clinical decision support is very important especially in such a wide-spread disease as a coronary artery disease. A large variety of prediction methods can potentially solve the classification problem to support clinical decisions. However, not all of them provide similar efficiency for the classification of patients with coronary artery disease. We have analyzed prediction the efficiency of classifiers (Ridge Classifier, XGB Classifier and Logistic Regression) depending on the number and combination of features. We have tested 24 sets of features on 4 classifiers to proof the hypothesis that using optimized features sets with a higher Pearson ratio results in more efficient classifiers than using all available data. |
|---|---|
| Notas: | Gesehen am 28.04.2020 |
| Descripción Física: | Online Resource |
| ISBN: | 9781614999751 |
| DOI: | 10.3233/978-1-61499-975-1-199 |