Text
ANALISIS SENTIMEN KOMENTAR VAKSINASI COVID-19 DI INSTAGRAM MENGGUNAKAN DEEP LEARNING XLNET
The Covid-19 vaccination activity is one of the most discussed things on social media. Comments on social media can be used to look at public sentiment on the Covid-19 vaccination. This study aims to perform sentiment analysis on Instagram comments about Covid-19 vaccination using XLNet and look at its performance. This study used Wikipedia corpus data and 2.000 Indonesian comments from Instagram. This study uses two software, namely training and test software. Training software is used to pre-train and fine-tune the XLNet model. Test software used for testing sentiment analysis using XLNet. The results of sentiment analysis on Instagram comments using XLNet get the best accuracy value on the model using epochs value: 7 and batch size: 16 in the fine-tuning process with accuracy: 74,25%, precision: 70,73%, recall: 82,75%, and f-measure: 76,27%. This study found that epoch value and batch size changes in the fine-tuning process affect the model performance. The epoch value addition and the selection of a smaller batch size value in the fine-tuning process almost always increase the model accuracy. The epochs value addition reduces accuracy in model with epochs value: 8 and batch size: 16, also in model with epochs value: 7 and batch size: 50. Meanwhile, the selection of a smaller batch size reduces accuracy in model with epochs value: 3 and batch size: 32 as well as a model with epochs value: 8 and batch size: 16.
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2207004222 | T79439 | T794392022 | Central Library (Referens) | Available but not for loan - Not for Loan |
No other version available