Application of Data Mining Techniques in Diagnosing Various Thyroid Ailments: A Review

The thyroid gland plays one of the most important organs in the human body. It secretes thyroid hormones, which regulate metabolism. Hypothyroidism and hyperthyroidism are caused by either too little or too much thyroid hormone secretion. This study assesses and analyzes existing data mining methods for diagnosing thyroid diseases. This paper aims to provide and identify the best practices in terms of applying data mining techniques such as decision tree, k-nearest neighbor, SVM, PNN, various Thyroid ailments which include the best machine learning model, naive Bayes, etc. Also, this research evaluates the preliminary techniques used to diagnose various thyroid diseases based on their efficacy and the number of attributes under the evaluation matrix. The attributes Age, sex, TSH, T3, TBG, T4U, TT4, and FTI were determined to be the most commonly used medical attributes in previous research works to perform experimental work to diagnose thyroid disorders. Almost every researcher has utilized one or more of these features to perform thyroid disease diagnostic work. According to the results of this study, there is a relationship between the number of attributes used and the accuracy rate achieved; The noticeable results that were presented in this study are some models are higher with fewer feature attributes while with the advent of the neural networks, the higher that number of attributes can give a better performance of classification. This area could be explored by considering adding and using more features to provide a more accurate and reliable output that can be a baseline for development.