Összesen 2 találat.


001-es BibID:BIBFORM075341
Első szerző:Sütő József (programtervező informatikus)
Cím:Music stimuli recognition from electroencephalogram signal with machine learning / Jozsef Suto, Stefan Oniga, Petrica Pop Sitar
Megjegyzések:When humans are listening to music they perceive beats, rhythms and melodies. This is the basis of music stimuli recognition where the goal is to explore how music influences our brain activity. In previous studies the emotional state determination was based on users' feedback. However this method is unreliable in most cases because an emotion state is not exact and it is changing relatively slowly. In this paper we tried to recognize music-induced electroencephalogram patterns from the well-known Neurosky Mindwave Mobile device's signal with feed forward artificial neural network. The paper describes our self-developed EEG measurement framework and the efficiency of the neural network with different kinds of feature extraction strategies.
Tárgyszavak:Műszaki tudományok Informatikai tudományok előadáskivonat
EEG signal
digital signal processing
feature extraction
machine learning
music stimuli
Megjelenés:2018 7th International Conference on Computers Communications and Control (ICCCC) : proceedings. - p. 260-264. -
További szerzők:Oniga István László (1960-) (villamosmérnök) Pop Sitar, Petrica (1972-) (matematikus)
Pályázati támogatás:ÚNKP-17-3-IV
Internet cím:Intézményi repozitóriumban (DEA) tárolt változat
Szerző által megadott URL


001-es BibID:BIBFORM065854
035-os BibID:(IEEE)7496749 (Scopus)84979986117
Első szerző:Sütő József (programtervező informatikus)
Cím:Comparison of wrapper and filter feature selection algorithms on human activity recognition / Jozsef Suto, Stefan Oniga, Petrica Pop Sitar
Megjegyzések:Feature selection is an increasingly important part ofmachine learning. The purpose of feature selection is dimensionreduction in a large multi-dimensional data set and it can be thekey step of successful knowledge discovery in those problemswhere the number of features is large. This research area hashuge practical significance because it accelerates decisions andimproves performance. The requirements of specific applicationsin different kinds of research areas have led to the developmentof new feature selection techniques with different properties. Inthe last few decades, several feature selection algorithms havebeen proposed with their particular advantages anddisadvantages. Despite of the intensive research and the largeamount of works, the different kinds of feature selectionalgorithms have not been tested yet in the human activityrecognition problem. It was the main motivation of our work andthis paper tries to fill this gap. Therefore, in this article wepresent a conceptually simple naïve Bayesian wrapper featureselection method and compare it with some widely used filterfeature selection algorithms. The result of this workdemonstrates that, the wrapper technique outperforms filteralgorithms in this type of problem. In addition, this paper showsan example, when the classifier dependency of a wrapper methoddo not visible
Tárgyszavak:Műszaki tudományok Informatikai tudományok tanulmány, értekezés
artificial neural network
feature selection
human activity recognition
machine learning
Megjelenés:2016 6th International Conference on Computers Communications & Control (ICCCC) : This edition of conference is dedicated to the 110th anniversary of Grigore C. Moisil (1906-1973) Romanian Mathematician, Computer Pioneer Award of IEEE Computer Society (1996 - posthumously) / eds. I. Dzitac, F.G. Filip, M.J. Manolescu. - p. 124-129. -
További szerzők:Oniga István László (1960-) (villamosmérnök) Pop Sitar, Petrica (1972-) (matematikus)
Internet cím:Szerző által megadott URL
Intézményi repozitóriumban (DEA) tárolt változat
Rekordok letöltése1