Web-Based Therapy's Exercise Generator Expert System Design and Implementation Using Depth-First Search, Breadth-First Search, and Transitive Reduction Algorithms

• Main Author: Adzaka (NIM: 18213001), Fikriansyah
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/22161
• Institution: Institut Teknologi Bandung

Problems which will be solved in this final project are about how to design and implement a web-based expert system which can arrange therapy’s exercise and select which appropriate therapy module is given to appropriate patient and help therapist to assess currently running therapy. Also, this final project will solve how depth-first search, breadth-first search, and transitive reduction are being implemented corresponding with how therapist solve the problems in the context of giving and checking appropriate exercise for certain patient. <br /> <br /> Research start with analysis stage to understand about current system condition. In this stage, problems with selecting appropriate module to be checked when observation is being conducted, and when regression checking procedure is being conducted, were found. Then, those analysis about current system were break down into user requirements, functional requirements, and a nonfunctional requirement. The new therapy system after the expert system is implemented was also described. <br /> <br /> The next stage of the research is designing the expert system. In this stage, use case was defined for two user level, admin and therapist. After the use case was defined, flowcharts to model the algorithms were defined and Entity Relationship Diagram was defined to model the database. <br /> <br /> After all design had defined, the next stage of the research is implement the expert system and test it. Expert judgement approach was selected by selecting four therapists as samples, and the test was conducted by using blackbox method to evaluate the functionality of the system. The samples then give judgement wether the system can solve current problem or not. The result of the test shown that all functional requirements had fulfilled. All samples agreed that the system is easy to use and the system can help therapists to solve related problems. <br /> <br /> Conclusions of this research are, expert system had been successfully designed and implemented in the form of prototype of 80 modules, with all test samples tend to agree that the system is easy to use. The samples also tend to agree that the system can help therapist to select appropriate exercise for appropriate patient, can help therapist to assess patient ability, and can help therapist to fix the curriculum in the future. The depth-first search, breadth-first search, and transitive reduction algorithm have developed and can handle 80 modules well. This conclusion is drawn from the functional test result which all the test cases have passed for 80 modules. Also, samples tend to agree that the system correspond with therapy’s needs, which leads to the conclusion that the algorithm correspond with how therapy solve the problems related to give appropriate exercise to appropriate patient and guide therapist on checking patient. The recommendations for next related research are, learn how expert solve the problems first before anything else, and select appropriate inference algorithm correspond with how the expert solve related problems.