ISSN (Print): 2328-3491, ISSN (Online): 2328-3580, ISSN (CD-ROM): 2328-3629 American International Journal of Research in Science, Technology, Engineering & Mathematics AIJRSTEM 16-260; © 2016, AIJRSTEM All Rights Reserved Page 217 Available online at http://www.iasir.net AIJRSTEM is a refereed, indexed, peer-reviewed, multidisciplinary and open access journal published by International Association of Scientific Innovation and Research (IASIR), USA (An Association Unifying the Sciences, Engineering, and Applied Research) Thyroid Disease Diagnosis Using Hybrid Intelligent Systems Shaymaa R. Saleh 1 , Zamen F. Jabr 2 , , Wijdan R. Abdulhussein 3 1,3 University of Thi-Qar, College of Computers and Mathematics Sciences, Al-Nasiriyah, IRAQ. 2 University of Thi-Qar, College of Education for Girls, Al-Nasiriyah, IRAQ. I. Introduction Thyroid hormones produced by the thyroid gland helps control the body’s metabolism. The thyroid gland produces two active thyroid hormones, levothyroxine (abbreviated T4) and triiodothyronine (abbreviated T3). These hor- mones are important in the production of proteins, in the regulation of body temperature, and in overall energy production and regulation. The seriousness of thyroid disorders should not be underestimated as thyroid storm (an episode of severe hyperthyroidism) and myxedema coma (the end stage of untreated hypothyroidism) may lead to death in a significant number of cases [1-3]. Thyroid performance has a serious effect on many basic organs of the body. If thyroid disorder is not recognized on time, the patient will suffer from thyroid attack or coma which might lead to his death. Therefore, true diagnosis of thyroid disorder (Thyroid low efficiency and high efficiency) based on laboratory and clinical tests (disease symptom) are quite vital. The use of neural networks as a smart tool for classifying thyroid data is more accurate and flexible than other approaches. Today, the use of smart methods in control systems, signal processing and sample recognition is a powerful tool in various scientific, technical and engineering, medical, and medical engineering researches has been greatly emphasized. An example of the use of such systems in different medical and medical engineering fields could be the following: research on various diseases, simulation of various body organs, and simulation of body metabo- lisms. In this regard, the research on thyroid disease and its diagnosis by means of smart neural methods has been emphasized by researchers. In neural network approach, it has been tried to pattern human’s brain functions and nervous system. This method is capable of solving complicated issues by relying on learning capacities and par- allel processing in natural neural networks. Neural network capabilities and their application in different issues such as signal processing, sample recognition, patterning, identification, prediction, controlling, and optimization have been emphasized in recent decades and these structures are recently used with regard to their learning capa- bility as a common method which is independent of proposed model [4-5]. In general, thyroid disease can be divided into two broad groups of disorders: those, which primarily affect the function of the thyroid gland and those, which involve neoplasms, or tumors, of the thyroid. Both types of disor- ders are relatively common in the general population. Most thyroid problems can be treated successfully. Abnor- malities of thyroid function are usually related to production of too little thyroid hormone (hypothyroidism) or production of too much thyroid hormone (hyperthyroidism)[1-3]. In this study, thyroid disease diagnosis were realized by using hybrid intelligent system which is neuro fuzzy system. In Neuro Fuzzy system the neural network and fuzzy logic was combined to get the main features of artificial neural networks with those of fuzzy logic and to overcome some of the limitations of these techniques where neural network provides learning capability but cannot extract the knowledge from the connection weight and fuzzy logic provide the Interpretability by using if-then rules to describe the problem but cannot automatically Abstract: Diagnosis is important issues in patient human treatment because it potency and accuracy in deter- mining whether or not a patient encompasses a specific illness. There has been an outsized increase within the range of thyroid cases over the past few years. Diagnosis of thyroid disease is one among the necessary prob- lems to develop a medical decision support system which is able to facilitate the physicians to require effective decisions. In this paper, we present diagnosis system based on hybrid intelligent systems (Nero-fuzzy network) as classifier tool. In this paper the neural network and fuzzy logic was combined to get the main features of artificial neural networks with those of fuzzy logic and to overcome some of the limitations of these techniques. The experimental results presented for different proportions of training/testing groups show a high classifica- tion accuracy and convergence in rates. The overall accuracy is 100% for training and in range between 87 % and 95% for testing. Thyroid disease datasets are taken from UCI machine learning dataset. Key words: Thyroid disease, Diagnosis techniques, Neural network, Fuzzy logic, Nero-fuzzy classification