22 Ekim 2022 Konferans bildirisi Açık Erişim

İsmail Buğra Bölükbaşı; Betül Yağmahan

Citation Style Language JSON

{
  "ISBN": "987-625-8246-29-2", 
  "URL": "https://aperta.ulakbim.gov.tr/record/286136", 
  "abstract": "<p>In recent years, the number of people with diabetes has been increasing daily. Diabetes is an important<br>\ndisease that can cause serious damage to the body in the future and even cause death if precautions are<br>\nnot taken. Early and accurate detection of ever-increasing diabetes is gaining more importance in the<br>\nmedical world. The number of studies using machine learning methods to diagnose diabetes has<br>\nincreased significantly in the literature.<br>\nIn this study, type-2 diabetes disease was classified using different data preprocessing and machine<br>\nlearning methods on real-world data taken from a public hospital in Turkey. Logistic regression, Naive<br>\nBayes, C4.5, and Random Forest classification models were used in the study. In the classification<br>\nmodels, the patient&#39;s age, gender, complete blood count, biochemistry, and hormone test results were<br>\nused as input variables, and the disease diagnosis made by specialist doctors was used as output variable.<br>\nIn total, 43 different variables were studied. When the dataset was examined, it was noticed that there<br>\nwas an imbalance between the classes in the target variable. In cases where there is a class imbalance,<br>\nthe classification models can make incorrect assignments to the classes. To eliminate the class imbalance<br>\nin the data set used in the study, three different resampling methods were used: random undersampling<br>\n(RUS), random oversampling (ROS), and synthetic minority oversampling (SMOTE).<br>\nThe performances of four different machine learning methods were compared on each of the original<br>\ntraining dataset, random undersampled training dataset, random oversampled training dataset, and<br>\nsynthetic minority oversampled training dataset. A total of 16 different scenarios were studied.<br>\nAs a result of the analysis of all scenarios, four combinations that give the best results were determined.<br>\nThese are Naive Bayes working with original training dataset, Random Forest working with random<br>\nundersampled training and synthetic minority oversampled training datasets, and C4.5 algorithm<br>\nworking with random oversampled training dataset. The algorithm that takes the first place among the<br>\nfour scenarios that show the best results is the Random Forest algorithm working with random<br>\nundersampled training dataset.</p>", 
  "author": [
    {
      "family": "\u0130smail Bu\u011fra B\u00f6l\u00fckba\u015f\u0131"
    }, 
    {
      "family": "Bet\u00fcl Ya\u011fmahan"
    }
  ], 
  "container_title": "ABSTRACT BOOK", 
  "id": "286136", 
  "issued": {
    "date-parts": [
      [
        2022, 
        10, 
        22
      ]
    ]
  }, 
  "page": "330-331", 
  "publisher": "IKSAD Publishing", 
  "publisher_place": "Adana", 
  "title": "DIAGNOSIS OF DIABETES DISEASE USING MACHINE LEARNING METHODS IN AN IMBALANCED DIABETES DATASET", 
  "type": "paper-conference"
}