Development of medical decision support system for lower urinary tract dysfunction by classifying uroflowmetry-electromyography signals

Lower Urinary Tract Dysfunction (LUTD) is a disease which is commonly seen among children. Frequent urination, urinary incontinence, feeling of tightness and poor urine flow, etc., are main symptoms of LUTD. There are different ways for diagnosis of LUTD. The most popular method is Uroflowmetry-Electromyography (UF-EMG) test in which both UF and EMG signals are recorded synchronously. The main purpose of this study is to classify UF-EMG signals with different classification methods and determine which method will show the best classification performance. These results are expected to provide physicians supplementary information for diagnosis and treatment. Three different classification methods were used for this study. These methods are Support Vector Machine (SVM), Artificial Neural Networks (ANN) and K-Nearest Neighbor (KNN). In this study UF-EMG signals are analyzed and the outputs of SVM, ANN and KNN are classified into six different groups. All dataset has been analyzed by specialists and results of the learning algorithms are calculated considering specialist judgment as gold standard and found that 'Multi Layer Feed Forward Artificial Neural Network' structure has showed best classification performance.