Corona is a very contagious virus. In a pandemic like this, people often worry whether they are infected or not. When they cough, they often worry whether it is a sign of covid-19 or an ordinary cough. From the clinical symptoms can actually be known whether someone has Covid or not. In this study, a clinical symptom dataset will be used to classify the symptoms using a Decision Tree algorithm. The decision trees used in this research are J48 and Hoeffding Tree. Decision Tree is one of the most popular classification methods because it is easy to interpret by Humans. the prediction model uses a hierarchical structure. The concept is to convert data into decision trees or decision rules. the result of J48 were slightly better than the Hoeffding tree in terms of accuracy, precision, and recall. Meanwhile, from the tree view results, the Hoeffding Tree is simpler and the number of nodes is less than J48.
Deep learning semakin berkembang pesat dan banyak dimanfaatkan dalam berbagai bidang kehidupan. Salah satunya bisa dimanfaatkan untuk klasifikasi image medis penderita covid. Keras adalah salah satu framework deep learning yang paling banyak digunakan. Dalam Keras, terdapat beberapa macam algoritma optimizer. Salah satunya adalah optimizer Adam. Untuk menggunakan optimizer Adam ini, perlu menentukan angka learning rate. Penentuan angka learning rate sangat penting karena salah dalam menentukan angka learning rate akan berdampak pada hasil deep learning yang dilakukan. Batch size juga salah satu hyperparameter penting dalam deep learning. Penelitian ini bertujuan untuk mengetahui dan membandingkan beberapa learning rate dan batch size agar diketahui efek dan dampaknya pada hasil loss dan akurasi training dan validasi pada proses deep learning yang dilakukan. Ada 6 learning rate dan 3 batch size yang akan dibandingkan. Hasil yang optimal diantara 6 learning rate dalam penelitian ini adalah 0.0001 dan 0.00001. Sedangkan batch size yang paling bagus hasilnya dari tiga angka yang dibandingkan adalah batch size 5
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.