In this paper, summary of many experiments for various hybrid methods will be presented. Hybrid methods combine various methodologies from Data Mining such as data reduction, multiple classifiers systems, feature selection with rough sets methods. In the paper, three algorithms will be presented which will use the notion of surroundings and a k-NN method for data reduction. The paper also describes one multiple classifier system which uses several algorithms such as k-NN, decomposition trees and neural network. The rest of the paper focuses on five algorithms which use reducts and deterministic or inhibitory decision rules for feature selection. All the algorithms presented in the paper were tested on well known data sets from the UCI Repository of Machine Learning Databases. The algorithms presented in the paper have been implemented and can be tested in the DMES system. © Springer-Verlag Berlin Heidelberg 2013.
CITATION STYLE
Delimata, P., & Suraj, Z. (2013). Hybrid Methods in Data Classification and Reduction. Intelligent Systems Reference Library, 43, 263–291. https://doi.org/10.1007/978-3-642-30341-8_14
Mendeley helps you to discover research relevant for your work.