Automated Essay Assessment using Machine Learning: A Case Study on Newton's Laws of Motion

This study aims to develop and evaluate a machine learning–based essay assessment website designed to measure students’ understanding of Newton’s Laws of Motion. A Research and Development (R&D) approach was employed using the ADDIE model (Analysis, Design, Development, Implementation, and Evaluation). This research involved physics teachers in Jambi Province, with a total population of 60 teachers. A random sampling technique was applied to select 20 teachers as respondents. The product underwent two stages of expert validation: Stage 1 scored 3.51 (70.2%), categorized as feasible, while Stage 2 achieved 3.95 (79.04%), classified as highly feasible. Field evaluations conducted with teachers experienced in assessment reported a score of 4.45 (89%), indicating a very high level of practicality and usability. These findings demonstrate that the developed system is effective, reliable, and user-friendly, supporting teachers in providing deeper, constructive feedback and improving assessment efficiency. This research highlights the potential of machine learning integration in educational assessment and offers an innovative solution to enhance the quality of physics learning and school-based evaluation practice