Performance Evaluation of Automated Essay Scoring Online System for Competency Assessment of Community Academy

The essay scoring system has been done manually so far, which has become an obstacle in the learning evaluation process. The use of the Automated Essay Scoring (AES) system is an alternative to developing automated scoring applications. The purpose of this research paper is to evaluate the performance of the AES online system on the assessment of vocational education competencies in the IT Multimedia field at the Community Academy in Blitar and Pacitan, East Java. The method of this study is consisting of pre-processing text, similarity calculation with cosine similarity and correlation method to measure performance of AES with human rater. This study compares the performance of the assessment system using a human rater and the AES online system. The results show that the correlation value is 0.903569186 with excellent criteria and the MAE (Mean Absolute Error) value is 0.133547009 which indicates a fairly small error value.