science-review.ru
Scientific journal
Научное обозрение. Медицинские науки
ISSN 2500-0780
ПИ №ФС77-57452
ISSN 2500-0780
ПИ №ФС77-57452
ALGORITHM FOR THE PREDICTION AND PREVENTION OF STONE RECURRENCE IN UROLITHIASISWITH USING FUZZY LOGIC
259 KB
The article describes a method of fuzzy logic for prediction and prevention of urolithiasis. Formed dictionary informative features and alphabet of classes.developed the formula for calculating the membership functions according to the features, аnd using an iterative rule, calculate the coefficient of confidence in the examined object belonging to the desired class. By comparing the values obtained with the threshold values of the coefficients of confidence is defuzzification conclusion. Based on these decision rules,we developed an algorithm for selecting the prevention method in urolithiasis. The research of the effectiveness the synthesized decision rules was conducted. We made a prospective analysis of the stone recurrence after treatment 200 patients in Kursk Emergency Hospital. After one year of follow we had two groups of patient: 37 – with recurrent stone formation, 163 – without relapse. Using decision rules were calculated curative conclusions for these patientsand were constructed distribution histogram coefficient values. Analyze the intersection histogram shows a good agreement between the results of expert evaluation and synthesized decision rule in the control group: the diagnostic sensitivity – 0.9, the diagnostic specificity – 0.98, the positive predictive value – 0.98, the negative predictive value – 0.91 and diagnostic efficiency is 0.94.