Somdatta Patra and Gour Sundar Mitra Thakur, Lovely Professional University, India

The task of medical diagnosis with the help different intelligent system techniques is always crucial because it require high level of accuracy and less time consumption in decision making. Among all other AI techniques Artificial Neural Networks (ANN) as a tool for medical diagnosis has become the most popular in last few decades due to its flexibility and accuracy. ANN was developed after getting the inspiration from biological neurons. There are various diseases that are still needed to be diagnosed. Among many other critical diseases like cancer, thyroid disorder, diabetes, heart diseases, neuro diseases, asthma disease was also tried to bediagnosed effectively with various ANN mechanisms by different researchers. Due to various uncertainties about symptoms the study of Neuro-Fuzzy technique in this context became very popular in last few years. Neuro-Fuzzy now-a-days is one of the most advanced technique that is mainly concatenation of two model-neural networks and the fuzzy logic. In this model various parameters are used that are much crucial if ill-chosen and may led to failure of the whole system. Recent trend in analysis is following this model for advanced expert work. In this study an enhanced Neuro-fuzzy model has been proposed for the proper diagnosis of adult Asthma disease and to foster the proper aid or medication to the patients and make physicians alert for the upcoming disease pattern otherwise they may lack in the process of providing improper medication at right time. In the first phase data collected from various hospitals are used to train by three different types of learning of ANN like ANN with Self Organizing Maps (SOM), ANN with Learning Vector Quantization (LVQ) and ANN with Backpropagation Algorithm (BPA) through NF tool for much accurate result. In the second phase fuzzy rule base is applied to the classified data for the diagnosis of the disease.

Adult asthma, Artificial Neural Network (ANN), Neuro Fuzzy (NF), Learning Vector Quantization (LVQ), Self-Organizing Map (SOM),Backpropagation.