Classification of Chronic Obstructive Pulmonary Disease Based on Neuro-Fuzzy Software
This chapter presents a system for classification of chronic obstructive pulmonary disease (COPD) based on fuzzy rules and a trained neural network. Fuzzy rules and neural network parameters are defined according to Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines. For neural network training more than one thousand medical reports obtained from database of the company CareFusion were used. The system was subsequently validated in 285 patients by physicians at the Clinical Centre University of Sarajevo. Out of the investigated patients, 99.19% of the 248 with COPD and all of the 37 individuals with normal lung function were classified correctly. Obtained sensitivity (99.3%) and specificity (100%) in COPD were assessed, as well. Implemented neuro-fuzzy system for classification of COPD is based on a combination of spirometry and Impulse Oscillometry System (IOS) test results, which enables more accurate classification of the disease. Additionally, a complete patient’s dynamic assessment can be obtained rather than a mere static assessment through the use of bronchodilatation and bronchoprovocation.