Makine öğrenimi teknikleri kullanılarak parmak vuruş hareketi verileri ile parkinson hastalığı tespiti
Küçük Resim Yok
Tarih
2022
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Trakya Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Parkinson hastalığı dünyada en yaygın görülen ikinci nörodejeneratif (sinir hasarına bağlı geri dönüşü olmayan) rahatsızlıktır ve henüz kesin bir tedavisi bulunmamaktadır. Hastanın motor becerilerinde meydana gelen ve günlük aktivitelerini olumsuz yönde etkileyen hareket bozuklukları (yürüme, konuşma, denge vb.), uzman klinisyenler tarafından ilaç tedavileri uygulanarak azaltılmaya çalışılmaktadır. El hareketlerindeki bozulmalar (titreme, hareketlerde yavaşlama vb.) hastalığın ilk belirtileri arasında yer almakta olup, hastalığın erken evrede tespiti ile uygulanacak ilaç tedavileri maddi ve manevi açıdan büyük önem taşımaktadır. Bu amaçla, bu çalışmada Parkinson hastalığında klavye tuş vuruş dinamikleri 14 farklı veri kümesinde ilaç kullanım durumlarına, hastalık şiddetine ve cinsiyete göre incelenmiştir. Parkinson hastası ve sağlıklı katılımcılardan oluşan, donma, gecikme ve uçuş sürelerini içeren gün içerisinde arka planda çalışıp kayıt altına alınan ham verilerden öznitelikler çıkartılmış, makine öğrenmesi algoritmaları uygulanarak hangi klavye tuş vuruş özelliklerinin Parkinson hastalarını sağlıklı bireylerden ayırmada daha başarılı olduğu araştırılmıştır.
Parkinson's disease is the second most common neurodegenerative (irreversible due to nerve damage) disorder in the world and there is no definitive treatment yet. Movement disorders (gait, speech, balance, etc.), are tried to be reduced by the application of drug treatments by specialist clinicians which occur in the motor skills of the patient and negatively affect his daily activities. Impairments in hand movements (tremor, bradykinesia, etc.) are among the first symptoms of the disease and drug treatments to be applied with the detection of the disease at an early stage are of great importance in material and spiritual terms. For this purpose, in this study keyboard keystroke dynamics in Parkinson's disease were investigated in 14 different datasets according to drug use status, disease severity and gender. Features were extracted from the raw data including hold, latency and flight times recorded in the background during the day from Parkinson's patients and healthy participants and by applying machine learning algorithms, which keyboard keystroke characteristics were more successful in distinguishing Parkinson's patients from healthy individuals were investigated.
Parkinson's disease is the second most common neurodegenerative (irreversible due to nerve damage) disorder in the world and there is no definitive treatment yet. Movement disorders (gait, speech, balance, etc.), are tried to be reduced by the application of drug treatments by specialist clinicians which occur in the motor skills of the patient and negatively affect his daily activities. Impairments in hand movements (tremor, bradykinesia, etc.) are among the first symptoms of the disease and drug treatments to be applied with the detection of the disease at an early stage are of great importance in material and spiritual terms. For this purpose, in this study keyboard keystroke dynamics in Parkinson's disease were investigated in 14 different datasets according to drug use status, disease severity and gender. Features were extracted from the raw data including hold, latency and flight times recorded in the background during the day from Parkinson's patients and healthy participants and by applying machine learning algorithms, which keyboard keystroke characteristics were more successful in distinguishing Parkinson's patients from healthy individuals were investigated.
Açıklama
Yüksek Lisans
Anahtar Kelimeler
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control ; Elektrik ve Elektronik Mühendisliği
