Abstract:
In Ethiopia road traffic accidents are becoming big issue that needs critical solution by
government body with traffic police office to save life of the citizens. The purpose of this study is
to explore the possibility of developing road traffic accident control knowledge based expert
system (RTACKBE) prototype for making decision. RTACKBE system is a system that diagnoses
a road traffic accident by learning domain knowledge generated by expert and produce
preventive measurement as a solution. The system present mitigating measures and decide
potential alternatives solution for alleviation of traffic accidents by considering preventive
measurement to save life of citizen. This RTACKBES is scoped with vehicle to vehicle crash and
vehicle crash pedestrian. The source of data were from both primary (using such as observation
and interview); and secondary data sources like traffic accident report documents for design of
this study. To design system prototype researcher used KBS development methodology that
involves knowledge acquisition, representation and modeling steps. Moreover, domain
knowledge has been acquired from domain expert using data collecting method particularly by
interviewing and from data mining techniques to process past traffic accident record. Acquired
knowledge was represented by sets of rule determined by PART algorithm for extracted
knowledge from data mining. Acquired and represented knowledge was modeled based on
decision tree structure resulted by J48 prune algorithm. Finally Web development tools and
Waikato Environment for Knowledge Analysis (WEKA) data mining tools have been used for the
design of the system. The performance of system was evaluated and tested using two evaluation
performance; one: using test case and two: using user acceptance. As a result of finding the
prototype knowledge based expert system is promising and applicable in the domain area
(Jimma town) with the user acceptance of 84.17%. The performance of the system was achieved
with 84.2% due to perform and process expected task operation. The importance and
requirement of the system on domain area was evaluated by domain expert based on simplicity in
terms of its efficiency and ease of using was 97.5%. The user accepted RTACKBE system as a result of
user acceptance and test case evaluation and performance results are achieved the objectives of this
study.
