Kadarisman Kadarisman, Retno Kusumaningrum, Linggar Maretva Cendani, Selvi Fitria Khoerunnisa, Sukmawati Nur Endah, Khadijah Khadijah, Rismiyati Rismiyati, Priyo Sidik Sasongko, Afriani Afriani, Irdha Yusra
The Automated Essay Scoring (AES) application was designed to support the learning process and assessment of e-learning platforms. An essay question is a question that can capture students’ analytical and synthesis abilities. However, the assessment of this form of questions is still done manually, and it is not the flexible learning concept of e-learning. Therefore, this study aims to develop an application that automatically scores essay questions. The application is a web-based application with system development using an iterative waterfall model – the automatic scoring employed several deep learning models, i.e., IndoBERT as the pre-trained word embedding model and CNN-LSTM with Mean Over Time as the model for automated final scoring. Three implemented testing mechanisms are unit testing using black box testing, system integration testing and model performance testing based on QWK and loss value. All test cases that have been tested are indicated to be successful. In addition, the deep learning model performance results in a loss value of 0.04 and QWK of 0.451. © 2023 ICIC International.
Faculty of Education and Teacher Training, Universitas Terbuka, Jalan Cabe Raya, Pondok Cabe, Pamulang, Banten, Tangerang Selatan, 15437, Indonesia; Faculty of Law, Social and Political Science, Universitas Terbuka, Jalan Cabe Raya, Pondok Cabe, Pamulang, Banten, Tangerang Selatan, 15437, Indonesia; Department of Informatics, Faculty of Science and Mathematics, Universitas Diponegoro, Jl. Prof. Soedarto, SH Tembalang, Central Java, Semarang, 50275, Indonesia; Department of Management, Universitas Negeri Padang, Jalan Prof. Dr. Hamka. Rd, Air Tawar, West Sumatera, Padang, 25131, Indonesia
