User Guide
Why can I only view 3 results?
You can also view all results when you are connected from the network of member institutions only. For non-member institutions, we are opening a 1-month free trial version if institution officials apply.
So many results that aren't mine?
References in many bibliographies are sometimes referred to as "Surname, I", so the citations of academics whose Surname and initials are the same may occasionally interfere. This problem is often the case with citation indexes all over the world.
How can I see only citations to my article?
After searching the name of your article, you can see the references to the article you selected as soon as you click on the details section.
  Citation Number 3
 Views 18
 Downloands 1
Normalizasyon Yöntemlerinin Biyomedikal Verilerde Sınıflandırma Performansına Etkisi
2021
Journal:  
Avrupa Bilim ve Teknoloji Dergisi
Author:  
Abstract:

Günümüzde tıpta hastalıklarla ilgili veri miktarı artmakta ve bu verilerden hastalığın sınıfı hakkında tahminler yapılmaktadır. Bu tahminlere için olumlu sonuç sağlayabilecek teknikler üzerinde çalışmalar geliştirilmektedir. Bu tahminler için kullanılacak olan sınıflandırma algoritmaları bu tekniklerle daha doğru sınıflandırma başarısı gösterebilmektedir. Bu çalışmada karaciğer ve kalp hastalığı veri setleri ondalık ölçekleme, minimum maksimum, z-skor ve norm normalizasyon yöntemleriyle normalize edilmiştir. Daha sonra bu normalize edilen veriler ve ham verilere, 4 farklı k-kat çapraz değerlendirmede (2,5,10,20) yapay sinir ağları, Naive Bayes, destek vektör makinesi, k-NN ve karar ağacı gibi çeşitli sınıflandırma algoritmalarıyla ORANGE programı kullanılarak sınıflandırma işlemine tabi tutulmuştur. Verilerin sınıflama doğrulukları değerlendirilmiş ve normalizasyon yöntemlerinin sınıflandırma performansını artırdığı gözlemlenmiştir.

Keywords:

The Effect of Normalization Methods on Classification Performance in Biomedical Data
2021
Author:  
Abstract:

Today, the amount of data related to diseases in medicine is increasing, and from these data, predictions about the class of the disease are made. Studies are developed on techniques that can provide positive results for these predictions. Classification algorithms that will be used for these predictions can show success in more accurate classification with these techniques. In this study, data sets of liver and heart disease were normalized with decimal measurement, minimum maximum, z-score and norm normalization methods. Later, these normalized data and raw data were submitted to the classification process using the ORANGE program with various classification algorithms such as artificial nerve networks (2,5,10,20) in 4 different k-cat cross assessments, Naive Bayes, support vector machine, k-NN and decision tree. The classification accuracy of the data has been evaluated and the normalization methods have been observed to improve the classification performance.

Keywords:

Effect Of Normalization Techniques On Classification Success In Biomedical Data
2021
Author:  
Abstract:

Nowadays, the amount of data about diseases in medicine is increasing and predictions about the class of the disease are made from these data. Studies are being developed on techniques that can provide positive results for these predictions. Classification algorithms that will make these predictions can show more accurate classification success with these techniques. In this study, liver and heart disease data sets were normalized using decimal scaling, minimum maximum, z-score and norm normalization methods. Then, these normalized datas and raw datas are classified using the ORANGE program with various classification algorithms such as artificial neural networks, decision tree, support vector machine, k-NN and Naive Bayes in 4 different k-fold cross validation (2,5,10,20) has been processed. The classification accuracies of the data were evaluated and were observed that normalization methods increased the classification performance.

Keywords:

Citation Owners
Attention!
To view citations of publications, you must access Sobiad from a Member University Network. You can contact the Library and Documentation Department for our institution to become a member of Sobiad.
Off-Campus Access
If you are affiliated with a Sobiad Subscriber organization, you can use Login Panel for external access. You can easily sign up and log in with your corporate e-mail address.
Similar Articles








Avrupa Bilim ve Teknoloji Dergisi

Field :   Fen Bilimleri ve Matematik; Mühendislik

Journal Type :   Uluslararası

Metrics
Article : 3.175
Cite : 5.553
2023 Impact : 0.178
Avrupa Bilim ve Teknoloji Dergisi