Limits...
Ranking of characteristic features in combined wrapper approaches to selection.

Stańczyk U - Neural Comput Appl (2014)

Bottom Line: To estimate relevance of attributes and select their subset for a constructed classifier typically either a filter, wrapper, or an embedded approach, is implemented.Next, another predictor exploits this resulting ordering of features in their reduction.The proposed methodology is illustrated firstly for a binary classification task of authorship attribution from stylometric domain, and then for additional verification for a waveform dataset from UCI machine learning repository.

View Article: PubMed Central - PubMed

Affiliation: Institute of Informatics, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, Poland.

ABSTRACT

The performance of a classification system of any type can suffer from irrelevant or redundant data, contained in characteristic features that describe objects of the universe. To estimate relevance of attributes and select their subset for a constructed classifier typically either a filter, wrapper, or an embedded approach, is implemented. The paper presents a combined wrapper framework, where in a pre-processing step, a ranking of variables is established by a simple wrapper model employing sequential backward search procedure. Next, another predictor exploits this resulting ordering of features in their reduction. The proposed methodology is illustrated firstly for a binary classification task of authorship attribution from stylometric domain, and then for additional verification for a waveform dataset from UCI machine learning repository.

No MeSH data available.


ANN classification accuracy observed in sequential backward elimination process, in relation to the number of considered features, and for each average, there is indicated maximal and minimal performance
© Copyright Policy - OpenAccess
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4305374&req=5

Fig2: ANN classification accuracy observed in sequential backward elimination process, in relation to the number of considered features, and for each average, there is indicated maximal and minimal performance

Mentions: If such smaller network classifies not worse than before reduction, it means that the relevance of the recently discarded input is negligible and it can be treated as redundant. The performance is illustrated in Fig. 2, while Fig. 3 shows what happens to the classification accuracy of the system when the input features are reduced while following the reversed ANN Ranking. The two graphs from Figs. 2 and 3 show the same trends that are visible in the previously plotted performance of DRSA classifiers in Fig. 1.Fig. 2


Ranking of characteristic features in combined wrapper approaches to selection.

Stańczyk U - Neural Comput Appl (2014)

ANN classification accuracy observed in sequential backward elimination process, in relation to the number of considered features, and for each average, there is indicated maximal and minimal performance
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4305374&req=5

Fig2: ANN classification accuracy observed in sequential backward elimination process, in relation to the number of considered features, and for each average, there is indicated maximal and minimal performance
Mentions: If such smaller network classifies not worse than before reduction, it means that the relevance of the recently discarded input is negligible and it can be treated as redundant. The performance is illustrated in Fig. 2, while Fig. 3 shows what happens to the classification accuracy of the system when the input features are reduced while following the reversed ANN Ranking. The two graphs from Figs. 2 and 3 show the same trends that are visible in the previously plotted performance of DRSA classifiers in Fig. 1.Fig. 2

Bottom Line: To estimate relevance of attributes and select their subset for a constructed classifier typically either a filter, wrapper, or an embedded approach, is implemented.Next, another predictor exploits this resulting ordering of features in their reduction.The proposed methodology is illustrated firstly for a binary classification task of authorship attribution from stylometric domain, and then for additional verification for a waveform dataset from UCI machine learning repository.

View Article: PubMed Central - PubMed

Affiliation: Institute of Informatics, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, Poland.

ABSTRACT

The performance of a classification system of any type can suffer from irrelevant or redundant data, contained in characteristic features that describe objects of the universe. To estimate relevance of attributes and select their subset for a constructed classifier typically either a filter, wrapper, or an embedded approach, is implemented. The paper presents a combined wrapper framework, where in a pre-processing step, a ranking of variables is established by a simple wrapper model employing sequential backward search procedure. Next, another predictor exploits this resulting ordering of features in their reduction. The proposed methodology is illustrated firstly for a binary classification task of authorship attribution from stylometric domain, and then for additional verification for a waveform dataset from UCI machine learning repository.

No MeSH data available.