Abstract: Respiratory disease is a medical term that encompasses pathological conditions affecting the organs and tissues that make gas exchange possible in higher organisms, and includes conditions of the upper respiratory tract, trachea, bronchi, bronchioles, alveoli, pleura and pleural cavity, and the nerves and muscles of breathing. Respiratory diseases range from mild and self-limiting, such as the common cold, life-threatening entities like bacterial pneumonia, pulmonary embolism, and lung cancer.In Upcoming days current means of diagnosis are obtrusive and ill-suited for real time applications. The respiration disorder features classification was achieved through various classification techniques. However, Support Vector Machine (SVM) and Decision Tree Bagging (DTB) based classifier does not have better accuracy for respiration disease prediction and also for cloud services to have flexible capacity for both storage and signal processing. Artificial Neural Network (ANN) and K-Nearest Neighbor (KNN) classifiers are having high computational cost. And also the above mentioned classification techniques have high sensitivity, time complexity is high due to long training process, and also it uses only offline data loggers. So we propose effective classification techniques with convenient and low cost automatic diagnosis method that uses wearable MEMS sensor technology. This sensor technology detects different types of respiratory disorders by means of changes in diameter of chest wall during breathing and also their parameters are computed. Then features are extracted and selected by Correlation-based Feature Selection (CFS). These features are classified by using Learning Vector Quantization (LVQ) and Modified-Fuzzy min-max classifier using Compensatory Neurons(M-FMCN) to get more accuracy, to reduce sensitivity, to use online data loggers, to reduce time complexity and to reduce the computational cost compared with SVM, DTB, ANN and KNN. We also propose some effective and drug free Breathing Therapy which helps patients to recover from respiratory problems without medicines. Finally, the experimental results show that the effectiveness of the proposed classification techniques compared with the other classification techniques.
Keywords: Respiratory disease, Support Vector Machine(SVM), Decision Tree Bagging (DTB),Artificial Neural Network (ANN), K-Nearest Neighbor (KNN),Feature Selection, Feature Classification, Learning Vector Quantization (LVQ), Modified-Fuzzy min-max classifier using Compensatory Neurons(M-FMCN).