Oktay Yildiz
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Oktay Yildiz has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
- Sumeyra Kus Ordu
Gazi University Engineering Faculty, Department of Computer Engineering
3 shared publications - Fatih Emre Boran
Gazi Üniversitesi
2 shared publications - İlYas Özdoğan
Türk Standardları Enstitüsü
2 shared publications M. Ali Akcayol
Bilgisayar Mhendislii Blm, Gazi niversitesi, Ankara, Trkiye
2 shared publications- Semih Demirel
Department of Information Systems, Gazi University, Ankara, Turkey.
2 shared publications Alisettar Hüseynli
Bilgisayar Mhendislii Blm, Gazi niversitesi, Ankara, Trkiye
1 shared publication
Latest Publications
- 4 August 2026
- Published by Discover STM Publishing Ltd. in Saudi Journal of Emergency Medicine
- Vol. 7 (1S) , 79
Comparison of Machine Learning Methods and Scoring Systems in Predicting MACE Risk
Introduction Machine learning (ML) models analyze various risk predictors to accurately estimate patients' risk by understanding nonlinear relationships and interactions among multiple variables. By many trials ML models tested for prediction of mortality and MACE risk for chest pain patients within 30 days. Aims & Objectives In this study, we aimed to compare the prediction accuracies of HEART and TMACS scores and various ML methods for estimating the risk of mortality and MACE in patients presenting to the ED with chest pain. Material & Methods The total number was 888, of consecutive patients aged 18 and older, presented to the ED of Ege University Hospital, Izmir, Türkiye, with chest pain or symptoms indicative of ischemia, applying the exclusion criteria, 328 patients were excluded, resulting in a final sample size of 560 patients. The data of 560 patients were used for both training and testing through K-Fold Cross Validation. For evaluation of the dataset, several ML methods were applied, including K-Nearest Neighbor (K-NN), Naive Bayes, Logistic Regression (LR), Support Vector Machine (SVM), and Decision Trees. The performance of models developed using these five methods to calculate 30-day mortality and MACE risks. Results The highest AUC value for prediction of 30-day MACE belonged to the LR ML model with (0,860), and followed by SVM (0,840), which are higher than HEART score AUC value (0,827) and similar to T-MACS scores (0,851). Even for 30-day mortality prediction LR has the highest AUC value (0,950) (HEART: 0,929, T-MACS: 0,875) While HEART, K-NN and SCM have 100% recall for mortality, LR has a recall of 0,998, with its balanced structure and highest AUC value, LR seems to be most preferable ML algorithm by clinical use to predict 30-day mortality. Conclusion Our LR model showed better AUC values for 30-days mortality than HEART scores, and SVM model found to be better than all models and scoring systems to predict the 30-days MACE. This suggest that these models could potentially be integrated into electronic patient records to better assess the risk of mortality and MACE.
- 4 August 2026
- Published by Discover STM Publishing Ltd. in Saudi Journal of Emergency Medicine
- Vol. 7 (1S) , 79
Comparison of Machine Learning Methods and Scoring Systems in Predicting MACE Risk
Introduction Machine learning (ML) models analyze various risk predictors to accurately estimate patients' risk by understanding nonlinear relationships and interactions among multiple variables. By many trials ML models tested for prediction of mortality and MACE risk for chest pain patients within 30 days. Aims & Objectives In this study, we aimed to compare the prediction accuracies of HEART and TMACS scores and various ML methods for estimating the risk of mortality and MACE in patients presenting to the ED with chest pain. Material & Methods The total number was 888, of consecutive patients aged 18 and older, presented to the ED of Ege University Hospital, Izmir, Türkiye, with chest pain or symptoms indicative of ischemia, applying the exclusion criteria, 328 patients were excluded, resulting in a final sample size of 560 patients. The data of 560 patients were used for both training and testing through K-Fold Cross Validation. For evaluation of the dataset, several ML methods were applied, including K-Nearest Neighbor (K-NN), Naive Bayes, Logistic Regression (LR), Support Vector Machine (SVM), and Decision Trees. The performance of models developed using these five methods to calculate 30-day mortality and MACE risks. Results The highest AUC value for prediction of 30-day MACE belonged to the LR ML model with (0,860), and followed by SVM (0,840), which are higher than HEART score AUC value (0,827) and similar to T-MACS scores (0,851). Even for 30-day mortality prediction LR has the highest AUC value (0,950) (HEART: 0,929, T-MACS: 0,875) While HEART, K-NN and SCM have 100% recall for mortality, LR has a recall of 0,998, with its balanced structure and highest AUC value, LR seems to be most preferable ML algorithm by clinical use to predict 30-day mortality. Conclusion Our LR model showed better AUC values for 30-days mortality than HEART scores, and SVM model found to be better than all models and scoring systems to predict the 30-days MACE. This suggest that these models could potentially be integrated into electronic patient records to better assess the risk of mortality and MACE.
- Open Access
- 28 April 2026
- Published by Selçuk University Journal of Current Nursing Research in Güncel Hemşirelik Araştırmaları Dergisi
- Vol. 6 (1) , 21-32
Predicting Adolescents' Exposure to Inappropriate Content during Social Media Use Using Classification Algorithms: A Data Mining Approach
Aim: This study aimed to predict adolescents’ exposure to inappropriate social media content using data mining.Methods: This descriptive study was conducted using 231 data records. The data was obtained from the publicly available “Kaggle database”. A dataset containing information related to social media usage was used. The dataset included variables related to social media use. Multiple classification algorithms (Nearest Neighbor, Naive Bayes, Logistic Regression, Support Vector Machines, Decision Trees) were applied to the dataset. These models were applied comparatively to predict the effects of social media usage variables on individuals’ likelihood of exposure to inappropriate content. Their performance was evaluated using statistical metrics. This aimed to identify the methods providing the highest accuracy and reliability.Results: The effectiveness of each method in predicting inappropriate content was evaluated and the best performing model was determined. It was determined that decision trees showed the best performance in the prediction model after Nearest Neighbor. It was determined that the performances of Naive Bayes, Logistic Regression and Support Vector Machines models were equal and their performances were lower compared to other models.Conclusion: According to the results of the prediction model, it is predicted that the increase in the number of social media platforms used by adolescents in the 17-19 age range and the amount of time they spend on these platforms on a daily basis will directly increase the frequency of encountering inappropriate content. The model reveals that the increase in the level of social media use increases the likelihood of encountering unsupervised or unfiltered content.
- 28 April 2026
- Published by Selçuk University Journal of Current Nursing Research in Güncel Hemşirelik Araştırmaları Dergisi
- Vol. 6 (1) , 21-32
Open AccessPredicting Adolescents' Exposure to Inappropriate Content during Social Media Use Using Classification Algorithms: A Data Mining Approach
Aim: This study aimed to predict adolescents’ exposure to inappropriate social media content using data mining.Methods: This descriptive study was conducted using 231 data records. The data was obtained from the publicly available “Kaggle database”. A dataset containing information related to social media usage was used. The dataset included variables related to social media use. Multiple classification algorithms (Nearest Neighbor, Naive Bayes, Logistic Regression, Support Vector Machines, Decision Trees) were applied to the dataset. These models were applied comparatively to predict the effects of social media usage variables on individuals’ likelihood of exposure to inappropriate content. Their performance was evaluated using statistical metrics. This aimed to identify the methods providing the highest accuracy and reliability.Results: The effectiveness of each method in predicting inappropriate content was evaluated and the best performing model was determined. It was determined that decision trees showed the best performance in the prediction model after Nearest Neighbor. It was determined that the performances of Naive Bayes, Logistic Regression and Support Vector Machines models were equal and their performances were lower compared to other models.Conclusion: According to the results of the prediction model, it is predicted that the increase in the number of social media platforms used by adolescents in the 17-19 age range and the amount of time they spend on these platforms on a daily basis will directly increase the frequency of encountering inappropriate content. The model reveals that the increase in the level of social media use increases the likelihood of encountering unsupervised or unfiltered content.
- Open Access
- 4 February 2026
- Published by Springer Nature in Physical and Engineering Sciences in Medicine
- Vol. 49 (2) , 863-879
A new model based on multi-axis vision transformer for chondromalacia patella diagnosis in magnetic resonance scans
A degenerative disease of the patellofemoral joint cartilage, chondromalacia patella (CMP) often results in anterior knee discomfort and functional disability. Determining the best course of therapy and stopping the progression of the disease depend on an accurate and timely diagnosis. In this work, we provide a deep learning architecture based on transformers for the classification of chondromalacia patella using magnetic resonance imaging (MRI) data. We assessed transformer-based designs including Multi-Axis Vision Transformer (MaxViT), Vision Transformer (ViT), and Swin Transformer in addition to convolutional neural network (CNN) based models like Google Network (GoogLeNet), Residual Network 18 (ResNet18), and Mobile Network v2 (MobileNetV2). We evaluated the models’ ability to differentiate between cases of chondromalacia patella and normal cases. With an accuracy of 0.9817, precision of 0.9821, recall of 0.9817, and F1-score of 0.9818, Multi-Axis Vision Transformer outperformed all other models on the testing dataset.
- 4 February 2026
- Published by Springer Nature in Physical and Engineering Sciences in Medicine
- Vol. 49 (2) , 863-879
Open AccessA new model based on multi-axis vision transformer for chondromalacia patella diagnosis in magnetic resonance scans
A degenerative disease of the patellofemoral joint cartilage, chondromalacia patella (CMP) often results in anterior knee discomfort and functional disability. Determining the best course of therapy and stopping the progression of the disease depend on an accurate and timely diagnosis. In this work, we provide a deep learning architecture based on transformers for the classification of chondromalacia patella using magnetic resonance imaging (MRI) data. We assessed transformer-based designs including Multi-Axis Vision Transformer (MaxViT), Vision Transformer (ViT), and Swin Transformer in addition to convolutional neural network (CNN) based models like Google Network (GoogLeNet), Residual Network 18 (ResNet18), and Mobile Network v2 (MobileNetV2). We evaluated the models’ ability to differentiate between cases of chondromalacia patella and normal cases. With an accuracy of 0.9817, precision of 0.9821, recall of 0.9817, and F1-score of 0.9818, Multi-Axis Vision Transformer outperformed all other models on the testing dataset.
- Open Access
- 21 August 2025
- Published by Journal of the Faculty of Engineering and Architecture of Gazi University in Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi
- Vol. 40 (3) , 1659-1672
Aralıklı tip-2 bulanık c-ortalama ile zaman serilerinin bulanık dilsel özetlemesi: BIST100 örnek hisse uygulaması
Son yıllarda büyük ilgi gören dilsel özetleme, büyük verilerden insanların anlayabileceği çıkarımlar sağlamaktadır. Dilsel özetlemenin önemli bir aşaması, verilerin ne kadar doğru yansıtıldığını gösteren doğruluk derecesinin belirlenmesidir. Bu doğruluk derecelerinin hesaplanmasında literatürde eşit bölümleme yöntemi kullanılmaktadır fakat bu yöntem verilerin yoğunlaştığı aralıkları göz ardı etmektedir. Bu çalışmada, verilerin yoğun olarak yer aldığı aralıkları dikkate alarak bulanık kümeleri dağıtan Aralıklı Tip-2 Bulanık C Ortalama (AT2BCO) bölümleme yöntemi önerilmektedir. Önerilen yöntem ile aralıklı tip-2 bulanık kümeler oluşturulmuş, bulanık kardinalite tabanlı olasılık ve olabilirlik temelli yaklaşımlar kullanılarak dilsel özetlerin doğruluk dereceleri hesaplanmıştır. Önerilen yaklaşım adım adım açıklanmış ve literatürde kullanılan eşit bölümleme yöntemi ile sonuçları karşılaştırılmıştır. Önerilen AT2BCO bölümleme yöntemi, Borsa İstanbul'da (BIST) işlem gören üç hisse senedine ait son on yıllık dönemi kapsayan finansal zaman serilerine uygulanmıştır. Elde edilen sonuçlar, literatürde yer alan eşit bölümleme çalışmalarına göre daha kararlı sonuçlar ürettiğini göstermektedir.
- 21 August 2025
- Published by Journal of the Faculty of Engineering and Architecture of Gazi University in Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi
- Vol. 40 (3) , 1659-1672
Open AccessAralıklı tip-2 bulanık c-ortalama ile zaman serilerinin bulanık dilsel özetlemesi: BIST100 örnek hisse uygulaması
Son yıllarda büyük ilgi gören dilsel özetleme, büyük verilerden insanların anlayabileceği çıkarımlar sağlamaktadır. Dilsel özetlemenin önemli bir aşaması, verilerin ne kadar doğru yansıtıldığını gösteren doğruluk derecesinin belirlenmesidir. Bu doğruluk derecelerinin hesaplanmasında literatürde eşit bölümleme yöntemi kullanılmaktadır fakat bu yöntem verilerin yoğunlaştığı aralıkları göz ardı etmektedir. Bu çalışmada, verilerin yoğun olarak yer aldığı aralıkları dikkate alarak bulanık kümeleri dağıtan Aralıklı Tip-2 Bulanık C Ortalama (AT2BCO) bölümleme yöntemi önerilmektedir. Önerilen yöntem ile aralıklı tip-2 bulanık kümeler oluşturulmuş, bulanık kardinalite tabanlı olasılık ve olabilirlik temelli yaklaşımlar kullanılarak dilsel özetlerin doğruluk dereceleri hesaplanmıştır. Önerilen yaklaşım adım adım açıklanmış ve literatürde kullanılan eşit bölümleme yöntemi ile sonuçları karşılaştırılmıştır. Önerilen AT2BCO bölümleme yöntemi, Borsa İstanbul'da (BIST) işlem gören üç hisse senedine ait son on yıllık dönemi kapsayan finansal zaman serilerine uygulanmıştır. Elde edilen sonuçlar, literatürde yer alan eşit bölümleme çalışmalarına göre daha kararlı sonuçlar ürettiğini göstermektedir.
- 22 July 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 296
A hybrid fuzzy-CNN model with feature extraction from MACD-based trends for stock market movement forecasting
- 22 July 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 296
A hybrid fuzzy-CNN model with feature extraction from MACD-based trends for stock market movement forecasting
- Open Access
- 27 March 2025
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 28 (2) , 683-695
Comparative Analysis of Ad Click Behavior Prediction Using GAN-Augmented Data and Traditional Machine Learning Techniques
In e-commerce, predicting click-through rates (CTR) is crucial to anticipate user behavior. User historical data can be used to extract interests and enhance CTR prediction, leading to higher accuracy. In this study, a Generative Adversarial Network (GAN) has been used to tackle the issue of insufficient dataset for click-through rates. Furthermore, six different machine learning algorithms have been assessed in predicting ad click behavior. For the experimental study, we obtained user demographic and online activity data from Kaggle, along with a binary label indi-cating ad clicks. To enhance the model's performance, we employed a GAN for data augmenta-tion and generated additional training examples. We compared the machine-learning algorithm's outcomes with and without GAN-based data augmentation to evaluate its predicted accuracy. According to the findings, most algorithms have increased sensitivity and specificity after utilis-ing GAN to augment the data, indicating that the generated data has improved their ability to accurately distinguish positive and negative events. GAN-based data augmentation boosted all models to varying degrees, according to the findings.
- 27 March 2025
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 28 (2) , 683-695
Open AccessComparative Analysis of Ad Click Behavior Prediction Using GAN-Augmented Data and Traditional Machine Learning Techniques
In e-commerce, predicting click-through rates (CTR) is crucial to anticipate user behavior. User historical data can be used to extract interests and enhance CTR prediction, leading to higher accuracy. In this study, a Generative Adversarial Network (GAN) has been used to tackle the issue of insufficient dataset for click-through rates. Furthermore, six different machine learning algorithms have been assessed in predicting ad click behavior. For the experimental study, we obtained user demographic and online activity data from Kaggle, along with a binary label indi-cating ad clicks. To enhance the model's performance, we employed a GAN for data augmenta-tion and generated additional training examples. We compared the machine-learning algorithm's outcomes with and without GAN-based data augmentation to evaluate its predicted accuracy. According to the findings, most algorithms have increased sensitivity and specificity after utilis-ing GAN to augment the data, indicating that the generated data has improved their ability to accurately distinguish positive and negative events. GAN-based data augmentation boosted all models to varying degrees, according to the findings.
- Open Access
- 24 January 2025
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 28 (1) , 187-195
Yüzey Pürüzlülüğünün Makine Öğrenmesi ile Tahmin Edilmesi
CNC tornalama genellikle metal parçaları işlemek için kullanılır. Son yüzeyin kalitesi, işlemin kalitesini değerlendirilmesi ve işleme koşullarını belirlemesinde önemli bir parametredir. İşleme performansını optimize etmek için, işleme parametreleri ve yüzey pürüzlülüğü değerleri arasındaki karmaşık ilişkileri araştırmak ve tahmin etmek gerekir. Yapay Sinir Ağı (YSA) modelleri, karmaşık ilişkileri öğrenme ve tahmin etme yetenekleri nedeniyle kesme koşullarının yüzey pürüzlülüğü üzerindeki etkilerini araştırmak için kullanılabilir. Bu çalışmada, Çoklu Lineer Regresyon (ÇLR) ve YSA yöntemleri kullanılarak tornalama sonrası yüzey pürüzlülüğü (Ra) değerlerini tahmin etmek için tahmin modelleri geliştirilmiştir. Tahmin modellerini geliştirmek için işleme deneyleri yapılmıştır. Deneylerde kesme hızı (m/dak), kesme derinliği (mm) ve ilerleme hızı (mm/dev) olmak üzere üç kontrol faktörü seviyesi kullanılmıştır. Geliştirilen modeller deneysel ölçümlerle doğrulanmış ve performansları değerlendirilmiştir. YSA tahminlerinin gerçek değerlere göre %87,6 doğruluğa sahip olduğu, çoklu regresyon tahminlerinin ise %78,4 doğruluğa sahip olduğu hesaplanmıştır. Çalışma, YSA yönteminin MLR yönteminden daha yüksek bir performansa sahip olduğunu ve yüzey pürüzlülüğü değerlerini tahmin etmek için kullanılabileceğini göstermiştir.
- 24 January 2025
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 28 (1) , 187-195
Open AccessYüzey Pürüzlülüğünün Makine Öğrenmesi ile Tahmin Edilmesi
CNC tornalama genellikle metal parçaları işlemek için kullanılır. Son yüzeyin kalitesi, işlemin kalitesini değerlendirilmesi ve işleme koşullarını belirlemesinde önemli bir parametredir. İşleme performansını optimize etmek için, işleme parametreleri ve yüzey pürüzlülüğü değerleri arasındaki karmaşık ilişkileri araştırmak ve tahmin etmek gerekir. Yapay Sinir Ağı (YSA) modelleri, karmaşık ilişkileri öğrenme ve tahmin etme yetenekleri nedeniyle kesme koşullarının yüzey pürüzlülüğü üzerindeki etkilerini araştırmak için kullanılabilir. Bu çalışmada, Çoklu Lineer Regresyon (ÇLR) ve YSA yöntemleri kullanılarak tornalama sonrası yüzey pürüzlülüğü (Ra) değerlerini tahmin etmek için tahmin modelleri geliştirilmiştir. Tahmin modellerini geliştirmek için işleme deneyleri yapılmıştır. Deneylerde kesme hızı (m/dak), kesme derinliği (mm) ve ilerleme hızı (mm/dev) olmak üzere üç kontrol faktörü seviyesi kullanılmıştır. Geliştirilen modeller deneysel ölçümlerle doğrulanmış ve performansları değerlendirilmiştir. YSA tahminlerinin gerçek değerlere göre %87,6 doğruluğa sahip olduğu, çoklu regresyon tahminlerinin ise %78,4 doğruluğa sahip olduğu hesaplanmıştır. Çalışma, YSA yönteminin MLR yönteminden daha yüksek bir performansa sahip olduğunu ve yüzey pürüzlülüğü değerlerini tahmin etmek için kullanılabileceğini göstermiştir.
- 13 November 2024
- Published by Springer Nature in Food Analytical Methods
- Vol. 18 (2) , 229-244
VisDist-Net: A New Lightweight Model for Fruit Freshness Classification
- 13 November 2024
- Published by Springer Nature in Food Analytical Methods
- Vol. 18 (2) , 229-244
VisDist-Net: A New Lightweight Model for Fruit Freshness Classification
- Open Access
- 25 September 2024
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 27 (4) , 1317-1326
Evrişimli Sinir Ağı Kullanarak Dengesiz Doppler Radar Verisinde Hedef Tespiti
Kritik öneme sahip askeri ve sivil yerleşkelerin korunması geçmişte olduğu gibi günümüzde önemini korumaktadır. Bu amaçla, çeşitli sensörler barındıran sistemler geliştirilmektedir. Sensörlerin sağladığı verilerden bilginin elde edilmesi donanımların en verimli şekilde kullanılması açısından önemlidir. Radar sistemleri keşif, gözetleme ve tespit amacıyla sıklıkla kullanılmaktadır. Radar ile tespit edilen nesnelerin sınıflandırılması amacıyla kural tabanlı ve makine öğrenmesi tabanlı yöntemler mevcuttur. Makine öğrenmesi tabanlı yaklaşımlarda uzman görüşüne gerek kalmadan zaman içerisinde hedef nesnenin karakteristik özellikleri model tarafından öğrenilir. Bu sebeple bu yöntemler kural tabanlı yöntemlere nazaran daha avantajlıdır. Gerçekleştirilen bu çalışmada, dengesiz Doppler Radarı verisi üzerinde İHA’ların diğer nesnelerden ayırt edilmesi amacıyla hedef sınıflandırması yapılmıştır. Deneysel çalışmalarda en yüksek başarım SMOTE kullanılarak dengeli hale getirilen veri setinde elde edilmiş, önerilen CNN modeli ile %99,99 doğruluğa ulaşılmıştır.
- 25 September 2024
- Published by Politeknik Dergisi in Journal of Polytechnic
- Vol. 27 (4) , 1317-1326
Open AccessEvrişimli Sinir Ağı Kullanarak Dengesiz Doppler Radar Verisinde Hedef Tespiti
Kritik öneme sahip askeri ve sivil yerleşkelerin korunması geçmişte olduğu gibi günümüzde önemini korumaktadır. Bu amaçla, çeşitli sensörler barındıran sistemler geliştirilmektedir. Sensörlerin sağladığı verilerden bilginin elde edilmesi donanımların en verimli şekilde kullanılması açısından önemlidir. Radar sistemleri keşif, gözetleme ve tespit amacıyla sıklıkla kullanılmaktadır. Radar ile tespit edilen nesnelerin sınıflandırılması amacıyla kural tabanlı ve makine öğrenmesi tabanlı yöntemler mevcuttur. Makine öğrenmesi tabanlı yaklaşımlarda uzman görüşüne gerek kalmadan zaman içerisinde hedef nesnenin karakteristik özellikleri model tarafından öğrenilir. Bu sebeple bu yöntemler kural tabanlı yöntemlere nazaran daha avantajlıdır. Gerçekleştirilen bu çalışmada, dengesiz Doppler Radarı verisi üzerinde İHA’ların diğer nesnelerden ayırt edilmesi amacıyla hedef sınıflandırması yapılmıştır. Deneysel çalışmalarda en yüksek başarım SMOTE kullanılarak dengeli hale getirilen veri setinde elde edilmiş, önerilen CNN modeli ile %99,99 doğruluğa ulaşılmıştır.
- Open Access
- 27 February 2024
- Published by Poltekkes Kemenkes Surabaya in Journal of Electronics, Electromedical Engineering, and Medical Informatics
- Vol. 6 (2) , 116-124
Implementation of C5.0 Algorithm using Chi-Square Feature Selection for Early Detection of Hepatitis C Disease
Hepatitis C, a significant global health challenge, affects 71 million people worldwide, with severe complications such as cirrhosis and hepatocellular carcinoma. Despite its prevalence and availability in rapid diagnostic tests (RDTs), the need for accurate early detection methods remains critical. This research aims to enhance hepatitis C virus classification accuracy by integrating the C5.0 algorithm with Chi-Square feature selection, addressing the limitations of current diagnostic approaches and potentially reducing diagnostic errors. This research explores the development of a machine learning model for hepatitis C prediction, utilizing a publicly available dataset from Kaggle. It encompasses preprocessing techniques such as label encoding, handling missing values, normalization, feature selection, model development, and evaluation to ensure the model's efficacy and accuracy in diagnosing hepatitis C. The findings of this study reveal that implementing Chi-Square feature selection significantly enhances the effectiveness of machine learning algorithms. Specifically, the combination of the C5.0 algorithm and Chi-Square feature selection yielded a remarkable accuracy of 96.75%, surpassing previous research benchmarks. This highlights the potent synergy between advanced feature selection techniques and machine learning algorithms in improving diagnostic precision. The study conclusively demonstrates that machine learning is an effective tool for detecting hepatitis C, showcasing the potential to enhance diagnostic accuracy significantly. As a future recommendation, adopting AutoML is suggested to periodically automate the selection of the optimal algorithm, promising further improvements in detection capabilities.
- 27 February 2024
- Published by Poltekkes Kemenkes Surabaya in Journal of Electronics, Electromedical Engineering, and Medical Informatics
- Vol. 6 (2) , 116-124
Open AccessImplementation of C5.0 Algorithm using Chi-Square Feature Selection for Early Detection of Hepatitis C Disease
Hepatitis C, a significant global health challenge, affects 71 million people worldwide, with severe complications such as cirrhosis and hepatocellular carcinoma. Despite its prevalence and availability in rapid diagnostic tests (RDTs), the need for accurate early detection methods remains critical. This research aims to enhance hepatitis C virus classification accuracy by integrating the C5.0 algorithm with Chi-Square feature selection, addressing the limitations of current diagnostic approaches and potentially reducing diagnostic errors. This research explores the development of a machine learning model for hepatitis C prediction, utilizing a publicly available dataset from Kaggle. It encompasses preprocessing techniques such as label encoding, handling missing values, normalization, feature selection, model development, and evaluation to ensure the model's efficacy and accuracy in diagnosing hepatitis C. The findings of this study reveal that implementing Chi-Square feature selection significantly enhances the effectiveness of machine learning algorithms. Specifically, the combination of the C5.0 algorithm and Chi-Square feature selection yielded a remarkable accuracy of 96.75%, surpassing previous research benchmarks. This highlights the potent synergy between advanced feature selection techniques and machine learning algorithms in improving diagnostic precision. The study conclusively demonstrates that machine learning is an effective tool for detecting hepatitis C, showcasing the potential to enhance diagnostic accuracy significantly. As a future recommendation, adopting AutoML is suggested to periodically automate the selection of the optimal algorithm, promising further improvements in detection capabilities.