Wyniki wyszukiwania - BazTech

1

Overview of feature selection methods used in malignant melanoma diagnostics

Michalska Magdalena

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2021

|

T. 11, nr 1

32--35

EN

Currently, a large number of trait selection methods are used. They are becoming more and more of interest among researchers. Some of the methods are of course used more frequently. The article describes the basics of selection-based algorithms. FS methods fall into three categories: filter wrappers, embedded methods. Particular attention was paid to finding examples of applications of the described methods in the diagnosis of skin melanoma.

PL

Obecnie stosuje się wiele metod selekcji cech. Cieszą się coraz większym zainteresowaniem badaczy. Oczywiście niektóre metody są stosowane częściej. W artykule zostały opisane podstawy działania algorytmów opartych na selekcji. Metody selekcji cech należące dzielą się na trzy kategorie: metody filtrowe, metody opakowujące, metody wbudowane. Zwrócono szczególnie uwagę na znalezienie przykładów zastosowań opisanych metod w diagnostyce czerniaka skóry.

2

A statistical comparison of feature selectiontechniques for solar energy forecastingbased on geographical data

EL Motaki Salou, El Fengour Abdelhak

Computer Assisted Methods in Engineering and Science

|

2021

|

Vol. 28, no. 2

105--118

EN

In recent years, solar energy forecasting has been increasingly embraced as a sustainablelow-energy solution to environmental awareness. It is a subject of interest to the scientificcommunity, and machine learning techniques have proven to be a powerful means toconstruct an automatic learning model for an accurate prediction. Along with the variousmachine learning and data mining utilities applied to solar energy prediction, the processof feature selection is becoming an ultimate requirement for improving model buildingefficiency. In this paper, we consider the feature selection (FS) approach potential. Weprovide a detailed taxonomy of various feature selection techniques and examine theirusability and ability to deal with a solar energy forecasting problem, given meteorologicaland geographical data. We focus on filter-based, wrapper-based, and embedded-basedfeature selection methods. We use the reduced number of selected features, stability, andregression accuracy and compare feature selection techniques. Moreover, the experimentalresults demonstrate how the feature selection methods studied can considerably improvethe prediction process and how the selected features vary by method, depending on thegiven data constraints.

3

A differential evolution approach to dimensionality reduction for classification needs

Martinović G., Bajer D., Zorić B.

International Journal of Applied Mathematics and Computer Science

|

2014

|

Vol. 24, no. 1

111--122

EN

The feature selection problem often occurs in pattern recognition and, more specifically, classification. Although these patterns could contain a large number of features, some of them could prove to be irrelevant, redundant or even detrimental to classification accuracy. Thus, it is important to remove these kinds of features, which in turn leads to problem dimensionality reduction and could eventually improve the classification accuracy. In this paper an approach to dimensionality reduction based on differential evolution which represents a wrapper and explores the solution space is presented. The solutions, subsets of the whole feature set, are evaluated using the k-nearest neighbour algorithm. High quality solutions found during execution of the differential evolution fill the archive. A final solution is obtained by conducting k-fold cross-validation on the archive solutions and selecting the best one. Experimental analysis is conducted on several standard test sets. The classification accuracy of the k-nearest neighbour algorithm using the full feature set and the accuracy of the same algorithm using only the subset provided by the proposed approach and some other optimization algorithms which were used as wrappers are compared. The analysis shows that the proposed approach successfully determines good feature subsets which may increase the classification accuracy.