Abstract:
Urinary tract infection is a common health problem that described by the existence of microbial
pathogens in any part of the urinary tract including the kidneys, ureters, bladder, or urethra.
Urinary tract infection is also a type of infectious disease that can occur in all groups of the
population. This study aimed to develop a knowledge-based system for the diagnosis and
treatment of urinary tract infection by collaborating data mining results with expert knowledge
by using design science research methodology. To extract the required knowledge, the researcher
collected 2016 instances of urinary tract infection dataset from Mizan Tepi University Teaching
Hospital and Bonga Hospital. In addition to documented data, the researcher has used a semi structured interview technique to acquire knowledge from domain experts by using the purposive
sampling technique. To develop the classifier model, the researcher has conducted 6 experiments
by using a partial decision tree, repeated incremental pruning, and decision tree algorithms. To
select the best classifier model, a model performance comparison was done, and then the best
performance result was achieved on the J48 algorithm under the 10-fold cross-validation test
option. The accuracy of the model was 92.16%. After the required knowledge was discovered
from the data mining and experts, it was combined into one knowledge base. Finally, the
knowledge-based system was developed by using Prolog and Java programming languages. In
order to provide test cases, the constructed system was assessed through user acceptability
testing and system performance testing. Lastly, the evaluation results showed 90% accuracy for
the system performance by using test cases and 94% for the user acceptance testing. As a result,
the developed system could support health workers to diagnose the types of infections and
recommend the proper treatments
