Analisis Sentimen Terhadap Pembelajaran Secara Daring Pasca Pandemi Covid-19 Menggunakan Metode IndoBERT

Abstrak

Abstract: The Covid-19 pandemic, which previously occurred in 2020-2022, made the Indonesian government change the learning method from face-to-face to online learning. This change forces students to be able to adapt to the online learning system, it raises various opinions from the community. Online learning is also carried out at Bina Insani University and the University wants to know the responses of its students regarding online learning. This is done by distributing a survey in the form of a google form to students and then the survey results will be reviewed by the University staff as input for future learning system improvements. Sentiment analysis is done to find out opinions or opinions are positive, negative or neutral. To classify a sentence, a method is needed that can classify a sentiment. The IndoBERT method is the method used in this research to get sentiment results with a testing method using Confusion Matrix which will calculate the accuracy value of the IndoBERT method. The test results conducted in this study resulted in a fairly high accuracy value of 87% with a precision value of 87%, recall of 91% and F1-score of 89%. Testing was done by testing 100 sentences with various sentiments.