DEVELOPMENT OF MOTORCYCLE FAULTATION DIAGNOSTIC FRAME-BASED EXPERT SYSTEM AND CHATBOT
A frame-based expert system is a program to reconstruct the expertise and reasoning abilities of experts in a limited domain using a knowledge base with frame representations. Until now, there has been no research that has developed a specific expert system for motorcycle diagnosis using frame-ba...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/69198 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | A frame-based expert system is a program to reconstruct the expertise and reasoning
abilities of experts in a limited domain using a knowledge base with frame
representations. Until now, there has been no research that has developed a specific
expert system for motorcycle diagnosis using frame-based representations. The
advantage of this representation is that it has the ability to represent domains that
have characteristics (stereotypes). So in this research, a frame-based expert system
is developed to diagnose the faultation that occurred.
The frame is designed using 16 classes that have stereotypical attributes of a
component. The knowledge base is formed by initiating and inheriting objects from
the class based on the acquisition of expert knowledge, then the problem solving
component matches the similarity and dissimilarity values between instances to
produce a solution. The expert system is built on a web environment utilizing a web
service that connects the chatbot as an interviewer component with problem solving
components that exist in the system. The input search process on the chatbot uses
the exact-match method based on the available key-values. The developed web
service consists of 3 APIs, each of which performs a keyword search process, a
hypothesis process, and provides assistance information.
Based on the process of collecting test cases with experts, there were 31 test cases
that actually happened. There were 51.61% (16 of 31 cases) defined in the existing
knowledge base. Some cases that are not in the base but exist in test cases, it is
concluded that there is additional information from experts at the time of knowledge
acquisition with the time of collecting test cases. Of the 16 test cases, 100% were
able to be correctly diagnosed by the expert system. |
|---|