Evolutionary Feature Subset Selection for Pattern Recognition Applications
نویسندگان
چکیده
A crucial part of a typical pattern recognition system is the extraction of the appropriate information that uniquely describes the patterns under processing. This information has the form of vectors and their contents are called features, which are constructed by specific extraction methods (Feature Extraction Methods FEMs). The length of the extracted feature vectors may take high dimension by incorporating many features for each pattern, although this huge information may be redundant and in a lot of cases this extra information corrupts the separability of the patterns under recognition. Therefore the need of an additional pre-processing method that reduces the feature vectors’ dimension, by selecting the most appropriate features, subject to some performance indices (class separability, high classification error etc.) is necessary. This procedure is called dimensionality reduction or feature subset selection and has attracted the attention of the scientific community for the last thirty years (Molina et al., 2002). This chapter is focused on the usage of evolutionary methods in selecting the appropriate feature subset from a pool of features, in a way the resulted subset increases the recognition rates in several benchmark pattern recognition problems. A simple genetic algorithm is used to examine the usefulness of a predefined feature set of some benchmark problems from the literature and some useful conclusions about the ability of these features to recognize the patterns are drawn. Moreover, the dependency of the resulted feature subsets, as far as their classification abilities are concerned, on the form of the fitness function used to measure the appropriateness of the candidate solutions, constructed by the genetic algorithm, is studied in this chapter. Three fitness functions with different properties are examined and their performance is compared to each other, for a set of pattern recognition problems.
منابع مشابه
Improving of Feature Selection in Speech Emotion Recognition Based-on Hybrid Evolutionary Algorithms
One of the important issues in speech emotion recognizing is selecting of appropriate feature sets in order to improve the detection rate and classification accuracy. In last studies researchers tried to select the appropriate features for classification by using the selecting and reducing the space of features methods, such as the Fisher and PCA. In this research, a hybrid evolutionary algorit...
متن کاملA New Hybrid Feature Subset Selection Algorithm for the Analysis of Ovarian Cancer Data Using Laser Mass Spectrum
Introduction: Amajor problem in the treatment of cancer is the lack of an appropriate method for the early diagnosis of the disease. The chemical reaction within an organ may be reflected in the form of proteomic patterns in the serum, sputum, or urine. Laser mass spectrometry is a valuable tool for extracting the proteomic patterns from biological samples. A major challenge in extracting such ...
متن کاملFeature subset selection using differential evolution and a wheel based search strategy
Differential evolution has started to attract a lot of attention as a powerful search method and has been successfully applied to a variety of applications including pattern recognition. One of the most important tasks in many pattern recognition systems is to find an informative subset of features that can effectively represent the underlying problem. Specifically, a large number of features c...
متن کاملEvolutionary Multi-Objective Feature Selection
Feature selection is one of the most pervasive problems in pattern recognition. It can be posed as a multiobjective optimisation problem, since, in the simplest case, it involves feature subset cardinality minimisation and performance maximisation. In many problem domains, such as in medical or engineering diagnosis, performance can more appropriately be assessed by ROC analysis, in terms of cl...
متن کاملA Hybrid Approach for Optimal Feature Selection based on Evolutionary Algorithms and Classic Approaches
Feature selection (FS) is a fundamental problem in the field of pattern recognition, which aims to find a minimal feature subset from the original feature space while retaining a suitably high accuracy in representing the original features. FS is used to improve the efficiency of learning algorithm especially for large scale datasets, by finding a minimal subset of features that has maximum eff...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2011