We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: BERT Fine-Tuning for Sentiment Analysis on Indonesian Mobile Apps Reviews

Abstract: User reviews have an essential role in the success of the developed mobile apps. User reviews in the textual form are unstructured data, creating a very high complexity when processed for sentiment analysis. Previous approaches that have been used often ignore the context of reviews. In addition, the relatively small data makes the model overfitting. A new approach, BERT, has been introduced as a transfer learning model with a pre-trained model that has previously been trained to have a better context representation. This study examines the effectiveness of fine-tuning BERT for sentiment analysis using two different pre-trained models. Besides the multilingual pre-trained model, we use the pre-trained model that only has been trained in Indonesian. The dataset used is Indonesian user reviews of the ten best apps in 2020 in Google Play sites. We also perform hyper-parameter tuning to find the optimum trained model. Two training data labeling approaches were also tested to determine the effectiveness of the model, which is score-based and lexicon-based. The experimental results show that pre-trained models trained in Indonesian have better average accuracy on lexicon-based data. The pre-trained Indonesian model highest accuracy is 84%, with 25 epochs and a training time of 24 minutes. These results are better than all of the machine learning and multilingual pre-trained models.
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Journal reference: SIET '21: 6th International Conference on Sustainable Information Engineering and Technology 2021
DOI: 10.1145/3479645.3479679
Cite as: arXiv:2107.06802 [cs.CL]
  (or arXiv:2107.06802v1 [cs.CL] for this version)

Submission history

From: Novanto Yudistira [view email]
[v1] Wed, 14 Jul 2021 16:00:15 GMT (876kb)

Link back to: arXiv, form interface, contact.