The Impact of Balancing Techniques and Feature Selection on Machine Learning Models for Diabetes Detection

Vahid Sinap

Article Details

The Impact of Balancing Techniques and Feature Selection on Machine Learning Models for Diabetes Detection

Authors:

Vahid Sinap

Publisher:

Fırat University

Main Topics:

Makine Öğrenme (Diğer)

Journal:

Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.37 Issue 1, pp.303-320

Document Type:

Article

Publication Date:

3/2025

Document Language:

English

Document ID (Org No.):

2001334873

pISSN:

1308-9072

Keywords:

Abstract:

 The detection of diabetes is crucial for effective management and prevention of the disease, which poses significant health risks globally. This study introduces a novel approach to diabetes detection by combining advanced data balancing techniques and feature selection methods, including Lasso L1 regularization, to enhance the performance of predictive models in imbalanced datasets. Techniques such as Random Under Sampling RUS, Adaptive Synthetic Sampling ADASYN, and Synthetic Minority Over-sampling Technique SMOTE were employed alongside models including Random Forest RF, CatBoost CB, Extreme Gradient Boosting XGB, K-Nearest Neighbors KNN, Gaussian Naive Bayes GNB, Logistic Regression LR, and Gradient Boosting GB to assess their impact on model accuracy and generalization capabilities. The findings reveal that the RF model achieved the highest accuracy of 93.25 when utilizing the SMOTE technique, underscoring the importance of appropriate data handling strategies in improving predictive outcomes. Furthermore, when all features were utilized without selection, the RF model attained an accuracy of 95.31, indicating the models capacity to capture complex patterns when feature richness is maximized. The comprehensive methodology used in the study achieved a higher accuracy in diabetes detection than research in the literature and provided important outputs for developing reliable prediction models in healthcare.

 The detection of diabetes is crucial for effective management and prevention of the disease, which poses significant health risks globally. This study introduces a novel approach to diabetes detection by combining advanced data balancing techniques and feature selection methods, including Lasso L1 regularization, to enhance the performance of predictive models in imbalanced datasets. Techniques such as Random Under Sampling RUS, Adaptive Synthetic Sampling ADASYN, and Synthetic Minority Over-sampling Technique SMOTE were employed alongside models including Random Forest RF, CatBoost CB, Extreme Gradient Boosting XGB, K-Nearest Neighbors KNN, Gaussian Naive Bayes GNB, Logistic Regression LR, and Gradient Boosting GB to assess their impact on model accuracy and generalization capabilities. The findings reveal that the RF model achieved the highest accuracy of 93.25 when utilizing the SMOTE technique, underscoring the importance of appropriate data handling strategies in improving predictive outcomes. Furthermore, when all features were utilized without selection, the RF model attained an accuracy of 95.31, indicating the models capacity to capture complex patterns when feature richness is maximized. The comprehensive methodology used in the study achieved a higher accuracy in diabetes detection than research in the literature and provided important outputs for developing reliable prediction models in healthcare.

 The detection of diabetes is crucial for effective management and prevention of the disease, which poses significant health risks globally. This study introduces a novel approach to diabetes detection by combining advanced data balancing techniques and feature selection methods, including Lasso L1 regularization, to enhance the performance of predictive models in imbalanced datasets. Techniques such as Random Under Sampling RUS, Adaptive Synthetic Sampling ADASYN, and Synthetic Minority Over-sampling Technique SMOTE were employed alongside models including Random Forest RF, CatBoost CB, Extreme Gradient Boosting XGB, K-Nearest Neighbors KNN, Gaussian Naive Bayes GNB, Logistic Regression LR, and Gradient Boosting GB to assess their impact on model accuracy and generalization capabilities. The findings reveal that the RF model achieved the highest accuracy of 93.25 when utilizing the SMOTE technique, underscoring the importance of appropriate data handling strategies in improving predictive outcomes. Furthermore, when all features were utilized without selection, the RF model attained an accuracy of 95.31, indicating the models capacity to capture complex patterns when feature richness is maximized. The comprehensive methodology used in the study achieved a higher accuracy in diabetes detection than research in the literature and provided important outputs for developing reliable prediction models in healthcare.

 Diyabet, küresel ölçekte önemli sağlık riskleri oluşturmaktadır. Diyabetin tespiti, hastalığın etkili yönetimi ve önlenmesi için büyük önem taşımaktadır. Bu çalışma, dengesiz veri setlerinde diyabet tespiti için çeşitli dengeleme tekniklerini ve Lasso L1 düzenlemesi de dahil olmak üzere özellik seçim yöntemlerini birleştirerek diyabet tespitine yeni bir yaklaşım getirmektedir. Çalışmada, Random Under Sampling RUS, Adaptive Synthetic Sampling ADASYN ve Synthetic Minority Over-sampling Technique SMOTE gibi teknikler, Random Forest RF, CatBoost CB, Extreme Gradient Boosting XGB, K-En Yakın Komşu KNN, Gaussian Naive Bayes GNB, Lojistik Regresyon LR ve Gradient Boosting GB modelleri ile kullanılarak bu tekniklerin model doğruluğu ve genelleme yetenekleri üzerindeki etkileri değerlendirilmiştir. Bulgular, SMOTE tekniği kullanıldığında RF modelinin 93,25 ile en yüksek doğruluğa ulaştığını göstermektedir, bu da uygun veri işleme stratejilerinin tahmin sonuçlarını iyileştirmede önemini vurgulamaktadır. Ayrıca, özellik seçimi yapılmaksızın tüm özellikler kullanıldığında, RF modeli 95,31 doğruluk elde etmiş ve bu da özellik zenginliği maksimize edildiğinde modelin karmaşık desenleri yakalama kapasitesini ortaya koymaktadır. Araştırmada kullanılan kapsamlı metodoloji, diyabet tespitinde literatürdeki araştırmalardan yüksek bir doğruluğa ulaşmış ve sağlık hizmetlerinde güvenilir tahmin modelleri geliştirmek için önemli çıktılar sağlamıştır.

Article Details

The Impact of Balancing Techniques and Feature Selection on Machine Learning Models for Diabetes Detection

Authors:

Vahid Sinap

Publisher:

Fırat University

Main Topics:

Makine Öğrenme (Diğer)

Journal:

Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.37 Issue 1, pp.303-320

Document Type:

Article

Publication Date:

3/2025

Document Language:

English

Document ID (Org No.):

2001334873

pISSN:

1308-9072

Diabetes detection

data balancing techniques

imbalanced datasets

predictive modeling

health informaticsDiyabet tespiti

veri dengeleme teknikleri

dengesiz veri setleri

tahmine dayalı modelleme

sağlık bilişimi

Issues

Article Details

Authors: Vahid Sinap

Publisher: Fırat University

Main Topics: Makine Öğrenme (Diğer)

Journal: Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.37 Issue 1, pp.303-320

Document Type: Article

Publication Date: 3/2025

Document Language: English

Document ID (Org No.): 2001334873

pISSN: 1308-9072

Issues

Authors:

Vahid Sinap

Publisher:

Fırat University

Main Topics:

Makine Öğrenme (Diğer)

Journal:

Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.37 Issue 1, pp.303-320

Document Type:

Article

Publication Date:

3/2025

Document Language:

English

Document ID (Org No.):

2001334873

pISSN:

1308-9072