A novel multi population based particle swarm optimization for feature selection

Kılıç, Fatih; Kaya, Yasin; Yıldırım, Serdar

doi:10.1016/j.knosys.2021.106894

Cited by 72 publications

(24 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Moreover, the number of the selected features was included in the evaluation function to get similar accuracy using a lower number of features. In this study, the fitness function as shown in equation 14 is performed [23], [57]- [59].…”

Section: Swarm-intelligence Based Feature Selectionmentioning

confidence: 99%

“…where ⊕ represents entry-wise multiplications, α > 0 means cuckoo's step size scaling parameter relating to the scales of the problem domain, β is Levy flight exponent. u and v are random numbers and calculated using equation (22)(23). x i represents position of i th egg (solution) and x g represents the best position in the current population.…”

Section: Swarm-intelligence Based Feature Selectionmentioning

confidence: 99%

“…In recent studies, swarm-intelligence algorithms have been widely used for engineering optimization problems [4], [19], global radiation forecasting [20], multiple unmanned aerial vehicles (UAV) control [21], and health monitoring of civil engineering structures [22]. Particle Swarm Optimization (PSO) [23], [24], Cuckoo Search (CS) [25], Grey Wolf Optimizer (GWO) [26], and Dragonfly Algorithm (DA) [27], [28] are the most widely used SI algorithms. PSO has been attracted researchers' attention because of its easy implementation and the small number of parameters to be adjusted.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Channel Selection Method for Emotion Recognition From EEG Based on Swarm-Intelligence Algorithms

2021

Self Cite

View full text Add to dashboard Cite

Increasing demand for human-computer interaction applications has escalated the need for automatic emotion recognition as emotions are essential for natural communication. There are various information sources that can be used for recognizing emotions, such as speech, facial expressions, body movements, and physiological signals. Among those physiological signals are more reliable for better affective communication with machines since they are almost impossible to control. Therefore, automatic emotion recognition from EEG signals has been a topic intensely investigated. Emotions are experiences that arise various cognitive functions observed in different frequency bands involving multiple brain areas and recognition from EEG with high accuracies is only possible with a large number of features extracted from the whole brain in various bands. Emotion regulation also requires integration of cognitive functions and thus functional connectivity between regions should also be considered. In this paper, we extract 736 features based on spectral power and phase-locking values. We particularly focus on finding salient features for emotion recognition using swarm-intelligence (SI) algorithms. We applied well-known classification algorithms for recognizing positive and negative emotions using the feature sets that are selected by these algorithms. Besides, features that are selected by all of them commonly are used as a new feature set. We report accuracies between 56.27% and 60.29% on the average; noting that by decreasing the feature size by 87.17% (from 736 to 94.40) an average accuracy of 60.01±8.93 was obtained with the random forest classifier. We also highlight the efficient electrode locations for emotion recognition. As a result, we define 11 channels as dominant and promising classification results are obtained.

show abstract

Section: Swarm-intelligence Based Feature Selectionmentioning

confidence: 99%

Section: Swarm-intelligence Based Feature Selectionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Channel Selection Method for Emotion Recognition From EEG Based on Swarm-Intelligence Algorithms

2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…The PSO algorithm was developed by Kennedy and Eberhart (1995) later many variants (Çomak, 2019; Kılıç et al, 2021; Zomorodi‐moghadam et al, 2021) are developed for various problem types. The main purpose of the initial study was to examine the social behaviour of birds and fish swarms and to ensure their graphical simulation.…”

Section: System Model For Automated Test Designmentioning

confidence: 99%

Automated test design using swarm and evolutionary intelligence algorithms

et al. 2021

Self Cite

View full text Add to dashboard Cite

The world's increasing dependence on computer‐assisted education systems has raised significant challenges about student assessment methods, such as automated test design. The exam questions should test the students' potential from various aspects, such as their intellectual and cognitive levels, which can be defined as attributes of the questions to assess student knowledge. Test design is challenging when various question attributes, such as category, learning outcomes, difficulty, and so forth, are considered with the exam constraints, such as exam difficulty and duration. In this paper, four contributions are provided to overcome test design challenges for the student assessment. First, a tool is developed to generate a synthetic question pool. Second, an objective function is designed based on the considered attributes. Third, the popular swarm and evolutionary optimization methods, namely particle swarm optimization, genetic algorithm, artificial bee colony, differential search algorithm are comparatively studied with novel methodologies applied to them. Finally, as the state of the art methods, artificial bee colony, and differential search algorithm are further modified to improve the solution of the test design. To perform the proposed algorithms, a dataset of 1000 questions is built with the proposed question attributes of the test design. Algorithms are evaluated in terms of their successes in both minimizing the objective function and running time. Additionally, Friedman's test and Wilcoxon rank‐sum statistical tests are applied to statistically compare the algorithms' performances. The results show that the improved artificial bee colony and the improved differential search provide better results than others in terms of optimization error and running time.

show abstract

“…Arı Kolonisi Algoritması (Karaboga ve Akay, 2007), Parçacık Sürü Algoritması (Eberhart ve Kennedy, y.y. ;Kılıç, Kaya ve Yildirim, 2021), Balina optimizasyon algoritması (Mirjalili ve Lewis, 2016), Gri Kurt Optimizasyonu (Aswani, Ghrera ve Chandra, 2016;Kumar, Chhabra ve Kumar, 2017;Mirjalili, Mirjalili ve Lewis, 2014), Karınca Kolonisi Algoritması (Dorigo, Birattari ve Stutzle, 2006), Ateşböceği algoritması (X. S. , Yarasa Algoritması (X.-S. ve Salp Sürü Algoritması (SSA) (Mirjalili ve diğerleri, 2017) doğadan esinlenen algoritmalardan bazılarıdır. Salp Sürü Algoritması basit, verimli ve esnektir, uygulanması kolaydır ve diğer algoritmalardan daha az sayıda parametreye sahiptir.…”

Section: Introductionunclassified

Salp Sürü Algoritması ile Öznitelik Seçimi ve Sınıflandırıcı Performans Değerlendirmesi

Can¹,

Kaya²,

Kılıç³

2021

European Journal of Science and Technology

Self Cite

View full text Add to dashboard Cite

ÖzSon yıllarda doğadan esinlenen sürü tabanlı algoritmalar arasında yer alan Salp Sürü Algoritması oldukça popüler olmuştur. Bu çalışmada, Salp Sürü Algoritması kullanılarak farklı veri setleri üzerinde öznitelik seçimi yapılmış, farklı sınıflandırıcılar ile bazı performans metrikleri karşılaştırılmıştır. Deneysel sonuçların hesaplanması için UCI Makine Öğrenmesi Deposunda yer alan BreastCancer, Colon ve Ionosphere veri setleri kullanılmıştır. Sınıflandırıcı olarak k En Yakın Komşu Algoritması, Destek Vektör Makineleri ve Rastgele Orman Algoritması kullanılmıştır. Sayısal sonuçlar incelendiğinde, çalışma zamanı bakımından kNN algoritması ile yapılan testler genellikle en hızlı algoritma olmuştur. Seçilen öznitelik sayısı bakımından ise SVM ve RF algoritmaları daha iyi sonuç vermiştir.

show abstract

A novel multi population based particle swarm optimization for feature selection

Cited by 72 publications

References 43 publications

A Channel Selection Method for Emotion Recognition From EEG Based on Swarm-Intelligence Algorithms

A Channel Selection Method for Emotion Recognition From EEG Based on Swarm-Intelligence Algorithms

Automated test design using swarm and evolutionary intelligence algorithms

Salp Sürü Algoritması ile Öznitelik Seçimi ve Sınıflandırıcı Performans Değerlendirmesi

Contact Info

Product

Resources

About