COMPARATIVE STUDY OF MACHINE LEARNING TECHNIQUES FOR SUPERVISED CLASSIFICATION OF BIOMEDICAL DATA

COMPARATIVE STUDY OF MACHINE LEARNING TECHNIQUES FOR SUPERVISED CLASSIFICATION OF BIOMEDICAL DATA

Článek recenzovaný mimo WoS a Scopus

Several classification methods have been widely used in literature for identification of diseases or differential diagnosis of various types of disorders. Classification methods such as support vector machines, random forests, AdaBoost, deep belief networks, K- nearest neighbors, linear discriminant analysis or perceptron are probably the most popular ones. Even if these methods are frequently used there is a lack of comparison between them to find better framework for classification. In this study, we compared performance of the above mentioned classification methods. The 10-fold cross validation was used to calculate accuracy and Matthews correlation coefficient of the classifiers. In each case these methods were applied to eight binary biomedical datasets. The same evaluation was realized also in conjunction with feature selection technique that passed only hundred most relevant features. Even though there is no single classification method that dominates in terms of performance, we found that some methods provide more consistent performance than others.

Several classification methods have been widely used in literature for identification of diseases or differential diagnosis of various types of disorders. Classification methods such as support vector machines, random forests, AdaBoost, deep belief networks, K- nearest neighbors, linear discriminant analysis or perceptron are probably the most popular ones. Even if these methods are frequently used there is a lack of comparison between them to find better framework for classification. In this study, we compared performance of the above mentioned classification methods. The 10-fold cross validation was used to calculate accuracy and Matthews correlation coefficient of the classifiers. In each case these methods were applied to eight binary biomedical datasets. The same evaluation was realized also in conjunction with feature selection technique that passed only hundred most relevant features. Even though there is no single classification method that dominates in terms of performance, we found that some methods provide more consistent performance than others.

SVM, AdaBoost, Random Forests, Deep Belief Networks, bioinformatics, microarray

SVM, AdaBoost, Random Forests, Deep Belief Networks, bioinformatics, microarray

DROTÁR, P.; SMÉKAL, Z.

30.09.2014

Technical University of Kosice

Kosice

1335-8243

Acta Electrotechnica et Informatica

2014

3

Slovenská republika

5

10

6

http://www.aei.tuke.sk/papers/2014/3/01_Drotar.pdf