Expert System For Improved Response In Greenfoot Game To The Machine Learning Implementation

Abstract

The game is the first semester material in the Algorithm course at the Pos Indonesia Polytechnic which was adopted from Oracle Academy. The output of this course is a simple 2D Greenfoot game. In 2020, around 70% of students made games, which move randomly. It is due to the computer responses that are not like human responses when the game is running and the rest uses a bouncing technique, so the computer response is very predictable. This study applied an expert system in the form of Naive Bayes and forward chaining with training data from student correspondence for the computer movements. The purpose of this study was to compare the random, bounce, and expert system techniques. This study contributes to providing concrete innovations in the form of course implementation in semester 1 (using Greenfoot) and semester 6 (using WEKA as a learning tool in artificial intelligence, data science, expert system courses) so that both are integrated. This can be a reference for teachers who use the Oracle Academy (Greenfoot) curriculum to be able to implement other machine learning/expert system models.