Floating search methods in feature selection
Pattern Recognition Letters
Selection of relevant features and examples in machine learning
Artificial Intelligence - Special issue on relevance
Wrappers for feature subset selection
Artificial Intelligence - Special issue on relevance
An introduction to variable and feature selection
The Journal of Machine Learning Research
Computers in Biology and Medicine
A Direct Method of Nonparametric Measurement Selection
IEEE Transactions on Computers
Multiclass support vector machines for diagnosis of erythemato-squamous diseases
Expert Systems with Applications: An International Journal
Expert Systems with Applications: An International Journal
Combined neural networks for diagnosis of erythemato-squamous diseases
Expert Systems with Applications: An International Journal
Feature selection with dynamic mutual information
Pattern Recognition
Expert Systems with Applications: An International Journal
Automatic Detection of Erythemato-Squamous Diseases Using k-Means Clustering
Journal of Medical Systems
Expert Systems with Applications: An International Journal
Hi-index | 0.00 |
This paper proposes hybrid feature selection algorithms to build the efficient diagnostic models based on a new accuracy criterion, generalized F-score (GF) and SVM. The hybrid algorithms adopt Sequential Forward Search (SFS), and Sequential Forward Floating Search (SFFS), and Sequential Backward Floating Search (SBFS), respectively, with SVM to accomplish hybrid feature selection with the new accuracy criterion to guide the procedure. We call them as modified GFSFS, GFSFFS and GFSBFS, respectively. These hybrid methods combine the advantages of filters and wrappers to select the optimal feature subset from the original feature set to build the efficient classifiers. To get the best and statistically meaningful classifiers, we not only conduct 10-fold cross validation experiments on training subset, but also on the whole erythemato-squamous diseases datasets. Experimental results show that our proposed hybrid methods construct efficient diagnosis classifiers with high average accuracy when compared with traditional algorithms.