An Intelligent System for Thyroid Disease Classification and Diagnosis

Abstract

Data mining Techniques play a vital role in healthcare organizations such as for decision making, diagnosing disease and giving better treatment to the patients. Thyroid gland plays a major role in maintaining the metabolism of human body. Data mining in health care industry provides a systematic use of the medical data. Thyroid diseases are most common today. Early changes in the thyroid gland will not affect the proper working of the gland. By the early identification of thyroid disorders, better treatment can be provided in the early stage thus can avoid thyroid replacement therapy and thyroid removal up to an extent. This paper proposes a method for the classification and diagnosis of thyroid disease that a user is suffering from along with disease description and healthy advices. Support Vector Machine is used for classification. To optimize SVM parameters Particle Swarm Optimization is applied. User is provided with a window to enter the details such as the values of TSH, T3, T4 etc. There may be some values missing while the user entering the values. K-Nearest Neighbor algorithm is used for approximating the missing values in the user input.