ISSN : 1796-203X
Volume : 1    Issue : 3    Date : June 2006

A Constructive Meta-Level Feature Selection Method based on Method Repositories
Hidenao Abe and Takahira Yamaguchi
Page(s): 20-26
Full Text:
PDF (369 KB)

Feature selection is one of key issues related with data pre-processing of classification task in a
data mining process. Although many efforts have been done to improve typical feature selection
algorithms (FSAs), such as filter methods and wrapper methods, it is hard for just one FSA to
manage its performances to various datasets. To above problems, we propose another way to
support feature selection procedure, constructing proper FSAs to each given dataset. Here is
discussed constructive metalevel feature selection that re-constructs proper FSAs with a method
repository every given datasets, de-composing representative FSAs into methods. After
implementing the constructive meta-level feature selection system, we show how constructive
meta-level feature selection goes well with 34 UCI common data sets, comparing with typical FSAs
their accuracies. As the result, our system shows the high performance on accuracies with lower
computational costs to construct a proper FSA to each given data set automatically.

Index Terms
Data Mining, Feature Selection, Constructive Meta-Processing.