Thy-Sys: A Preliminary Thyroid Wellness Assessment Through Machine Learning Using Pathological Factors

The thyroid gland is a butterfly-shaped organ located lower front of the neck that plays a critical role in one's overall well-being. According to survey, thyroid dysfunction is observed in 8.53% of Filipino adults aged 42 to 62 years old, implying that 1 out of 12 Filipino adults has some form of thyroid function abnormality. To promote awareness of thyroid health, the researchers developed the Thyroid System for Wellness Assessment (Thy-Sys) which aims to determine the wellness of a person's thyroid through machine learning. The system assesses the user's thyroid wellness based on their responses to questions related to observable characteristics of the more common thyroid diseases like hypothyroidism and hyperthyroidism. Prior to the development, four machine learning models, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Decision Tree, and SVM-KNN were evaluated in classifying the state of the thyroid through pathological factors. Training and testing of the machine were done over a dataset with 1,464 entries and a total of 18 pathological attributes. 10-fold cross-validation was performed to verify the models' accuracy while observing other metric scores. The SVM-KNN model garnered the highest accuracy of 99.55% with 98.75% precision, and 98.61 % recall and F1 scores, and was integrated into the system.