Sistem Berbasis Website Untuk Diagnosa Penyakit Hypertiroid Menggunakan Metode Artificial Neural Network

Abstract

Hyperthyroidism is a disease in which there are excessive levels of TT4 or T3, and sometimes both, but below normal levels of TSH, causing the body to experience cognitive decline, anxiety, and weight loss. The main cause of hyperthyroidism is Graves' disease. Graves' disease is an autoimmune disease that causes the body to produce excess thyroid hormones above normal limits. This non-communicable disease occurs more often in women than in men. Hormone indicators commonly used to diagnose hyperthyroidism are T3, TT4, and TSH levels. This study aims to diagnose hyperthyroidism using an artificial neural network (ANN) method with backpropagation algorithm, and this method will be used more accurately as decision-making advice. The model was trained and tested using the Kaggle dataset and data collection at a community health center, using a total of 700 pieces of data with 6 attributes and 1 class label. This model uses four different training and testing data distribution scenarios: 90:10, 80:20, 70:30, and 60:40. The model is evaluated by applying a confusion matrix consisting of precision, precision, recall, and f1 score. The best accuracy obtained in this study was 0.9 or 90% using a 60:40 dataset split and 7 hidden neurons, a learning rate of 0.01 and 800 epochs.