Skripsi

ANALISIS SENTIMEN ULASAN FILM INTERNET MOVIE DATABASE (IMDB) MENGGUNAKAN BIDIRECTIONAL ENCODER REPRESENTATIONS FROM TRANSFORMERS (BERT)

---

The large and unstructured volume of IMDb movie review data makes it difficult to map the opinions of the audience manually. Therefore, this study aims to develop a sentiment analysis system using BERT to classify IMDb movie reviews into positive and negative categories. The data used is secondary data amounting to 50,000 English-language data. This research stage includes pre-processing of text, data sharing with a ratio of 80:10:10 and processed through fine-tuning base-uncased BERT model.The training process was carried out through two fine-tuning experiments with different configurations. The first experiment used batch size 8, max length 128 without dropout and data augmentation, while the second experiment used batch size 32 and max length 256 by applying dropout and random token masking techniques for data augmentation. Each experiment was trained with six scenarios using a variation of learning rate of 2e-3 and 3e-5 and the number of epochs 3, 4 and 5. The results of this study show that the second experiment of scenario 5 with a learning rate configuration of 3e-5, epoch 4, batch size 32, and dropout 0.2 produced the best performance with an accuracy score of 93.26%, precision of 93.28%, recall of 93.26% and F1-score of 93.26%. Keywords: Sentiment analysis, IMDb, BERT, fine-tuning models

---

Availability

Detail Information

Series Title

-

Call Number

T2010632026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xix, 182 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

006.350 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Analisis Sentimen Data

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment