Feature Selection Techniques: A Review
Journal: International Journal for Scientific Research and Development | IJSRD (Vol.3, No. 11)Publication Date: 2016-02-01
Authors : Prajakta Kulkarni; S. M. Kamalapur;
Page : 690-693
Keywords : Feature Selection; Supervised; Semi-Supervised; Unsupervised;
Abstract
Feature is a prominent attribute of a process being observed. Set of features are applied to pattern recognition and machine learning algorithms for processing. In modern world, size of set of features has been increased to multiple of thousands. Hence dealing with large number of features became a challenge. Feature selection is one of the well-known technique to minimize the size of set of features. Feature selection is carried out mainly in three contexts: supervised, unsupervised and semi-supervised. Different measures are used for selection of features. In this paper some representative methods of each of the context are analysed along with their pros and cons. Also it gives idea about why there is a need of separate feature selection technique for semi-supervised data.
Other Latest Articles
- ANALYSIS ACCURACY AND PERFORMANCE OF DATA MINING TECHNIQUES IN HEALTHCARE
- DETERMINANTS OF THE NATIONAL ACHIEVEMENT TEST RESULTS OF THE SECONDARY SCHOOLS IN THE PROVINCE OF BILIRAN, PHILIPPINES
- THE S H E LL OF CLAM AS POTEN T IAL CATALYST RESOURCE FOR BIODIESEL PRODUCTION
- Development of An Integrated Website for Art Training
- A DRILLABILITY AND GEOMECHANICAL WELBORE STABILITY ANALYSIS FOR OPTIMAL BIT SELECTION - CASE STUDY OF A FIELD IN NIGER DELTA
Last modified: 2016-02-06 19:19:09