Ilham A. E. Zaeni, Dwiajeng Puspita Ratri, Agusta Rakhmat Taufani
Diarrhea is a global health issue with significant morbidity and fatality rates. Diarrhea is the second leading cause of mortality in children under the age of five, behind pneumonia. Many reasons can obstruct the diagnosis of diarrhea, including a lack of parental information about diarrhea, a continued underestimation of diarrhea, a lack of understanding about the kind of diarrhea, and misdiagnosis. The goal of this study is to identify and define the kind of diarrhea disease in toddlers, as well as to offer basic treatment options that can be conducted by the parents. The public health center of Mojopanggung provided the dataset for this study, which included 166 cases and 21 attributes. However, only eight attributes were shown have relation with the class following the Bivariate Pearson correlation test, namely defecate frequency, feces consistency, fever, nausea/vomiting, eye condition, thirst, and skin turgor. The categorization yielded a degree of dehydration, namely none, mild/moderate dehydration, and severe dehydration. Because the percentage of the number of classes in this dataset is still unbalanced, an oversampling approach is used to correct the problem, resulting in a final data count of 318 instances. The Naive Bayes Classifier (NBC) algorithm, which is a statistical classification approach for predicting the likelihood of membership in a class, is used in this work. Cross Validation was used in this study with a 10 folds test. for the categorization of kinds of diarrhea illnesses based on the degree of dehydration in toddlers, the findings of this study utilizing Confusion Matrix show an average value of 97.48 percent accuracy, precision of 96.55 percent, recall of 96.27 percent, and an error rate of 2.52 percent. © 2021 IEEE.
Universitas Negeri Malang, Dept. of Electrical Engineering, Malang, Indonesia
