Knowledge Discovery of Thyroid Diseases: Association Rule Mining Approach

Thyroid diseases are becoming a common, high-incidence condition across the globe. Hence, understanding the influencing variables is essential for accurate disease diagnosis. Many applications have been proposed by researchers to extract knowledge from the thyroid diseases dataset available in the UCI repository. Their objective was to improve the process by which medical professionals make diagnoses. This study applies Association Rule Mining (ARM) to investigate the hidden patterns in cases diagnosed with hyperthyroid and hypothyroid. Two rule generation algorithms (Apriori and FP-Growth) were used to explore the factors which contribute to thyroid conditions. The analysis showed that females have a high risk to contract hypothyroidism. The results also show that the condition is more likely to affect those over the age of 64 both males and females. On this basis, it is recommended to involve age and sex factors in the diagnosis guideline for medical experts to enhance the thyroid diagnosis process.