Open AccessPotential of Hybrid Adaptive Neuro Fuzzy Model in Simulating Clostridium Difficile Infection Status
Author(s) -
Ali Usman,
Ahmed Nouri Alsharksi,
Y.A Danmaraya,
Hadiza Usman Abdullahi,
Umar Muhammad Ghali
Publication year - 2020
Publication title -
international journal of basic sciences and applied computing
Language(s) - English
Resource type - Journals
ISSN - 2394-367X
DOI - 10.35940/ijbsac.a0191.073120
Subject(s) - context (archaeology) , clostridium difficile , adaptive neuro fuzzy inference system , medicine , logistic regression , receiver operating characteristic , reliability (semiconductor) , statistics , fuzzy logic , machine learning , computer science , artificial intelligence , mathematics , fuzzy control system , geography , power (physics) , physics , quantum mechanics , microbiology and biotechnology , biology , antibiotics , archaeology
The global burden posed by nosocomial diarrhea lead to the strong given attention by health practitioners science its morbidity and mortality rate hit about 500,000 rates annually in the United states. Diagnostic measures have been put in place to detect the presence of CD using different methods. Reliable prediction of the health status of patients is of paramount importance. This study aimed at investigating the status of stool samples collected to test the presence of clostridium difficile as either positive or negative from both inpatient and outpatient from the record units of Near East University Hospital using hybrid adaptive neuro fuzzy (known as ANFIS) model consisting of various combinations of membership functions and training Fis. In this context, the age of the patients, gender, results of the analysis conducted, the department in which the patient was admitted, the age category and the hospitalization were employed as the input variables. The performance accuracy of these membership functions and training FIS combinations were checked using two performance indices determination coefficient (R2) and mean square error (MSE). The obtained computation data driven models proves the reliability of the combination of subtractive clustering membership function and hybrid training FIS over the other three ANFIS combinations. Overall, the results indicated the reliability and satisfaction of hybrid adaptive neuro fuzzy in checking the status of stool samples collected to test the presence of clostridium difficile as either positive or negative from both inpatient and outpatient.