An automated content analysis model for forum moodle using semantic similarity

Abstract

The use of e-learning in Indonesia's educational environment is increasingly widespread. It requires an improvement in the method of evaluating learning outcomes in e-learning. Moodle is one of the most popular learning management systems in the world of education, but Moodle does not yet have the tools to evaluate learner activities in every Moodle activity. For example, Moodle doesn't have the tools to analyze the content of activity forums. This study proposes a model for analyzing message content in forums automatically. This model specifically aims to determine the relevance of the message to the course. The proposed model consists of three main processes: creating a dataset, detecting message relevance, and testing. Determination of the relevance of the message is based on the value of semantic similarity. Semantic similarity calculations are performed using Indonesian Wordnet and Levenshtein Distance Similarity. This model can help lecturers to detect which messages are relevant or worth and which messages are irrelevant or worthless toward the course. The proposed model is expected to improve the performance of evaluation systems of e-learning learning.