Publication detail

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

DROTÁR, P. SMÉKAL, Z.

Original Title

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

English Title

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

Type

journal article - other

Language

en

Original Abstract

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.

English abstract

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.

Keywords

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

RIV year

2014

Released

30.09.2014

Publisher

Technical University of Kosice

Location

Kosice

ISBN

1335-8243

Periodical

Acta Electrotechnica et Informatica

Year of study

2014

Number

3

State

SK

Pages from

5

Pages to

10

Pages count

6

URL

Documents

BibTex


@article{BUT112064,
  author="Peter {Drotár} and Zdeněk {Smékal}",
  title="COMPARATIVE STUDY OF MACHINE LEARNING TECHNIQUES FOR SUPERVISED CLASSIFICATION OF BIOMEDICAL DATA",
  annote="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.",
  address="Technical University of Kosice",
  chapter="112064",
  doi="10.15546/aeei-2014-0021",
  howpublished="print",
  institution="Technical University of Kosice",
  number="3",
  volume="2014",
  year="2014",
  month="september",
  pages="5--10",
  publisher="Technical University of Kosice",
  type="journal article - other"
}