Putra, Ferdyansyah Permana (2024) Implementasi IndoBERT Dalam Analisis Sentimen Berita Untuk Prediksi Harga Saham PT. Bank Rakyat Indonesia Tbk. Menggunakan Pendekatan Support Vector Regression. Diploma thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 2043201115_Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
2043201115_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (3MB) | Request a copy |
Abstract
Fluktuasi harga saham sering dipengaruhi oleh berbagai faktor, dan salah satunya adalah sentimen, yang dapat berasal dari masyarakat umum atau berita terkait saham. Harga saham PT. Bank Rakyat Indonesia Tbk. (BBRI), sebagai salah satu dari "big four" bank terbesar di Indonesia, juga dapat dipengaruhi oleh sentimen ini. Dalam usaha untuk meramalkan harga penutupan saham BBRI berdasarkan sentimen berita, penelitian ini mengadopsi pendekatan machine learning dengan menggunakan metode Support Vector Regression (SVR) dan mengoptimalkan fungsi dengan algoritma Fruit Fly Optimization Algorithm (FOA). Sentimen dievaluasi terlebih dahulu dengan metode IndoBERT. Penelitian ini mengusulkan beberapa model, termasuk model tanpa memperhitungkan sentimen, model dengan mempertimbangkan sentimen pada periode ke t, model dengan mempertimbangkan sentimen pada periode ke t-1, dan model dengan mempertimbangkan sentimen pada periode ke t dan periode ke t-1. Analisis hasil sentimen menggunakan IndoBERT menunjukkan tingkat akurasi sentimen secara keseluruhan di atas 90%. Selain itu, hasil pemodelan menunjukkan bahwa model terbaik menggunakan SVR adalah model yang tidak mempertimbangkan sentimen, dengan nilai error peramalan yaitu Mean Average Percentage Error (MAPE) lebih rendah dibandingkan dengan model lain. Hasil peramalan dengan dari model terbaik menunjukkan bahwa data aktual masih berada dalam interval kepercayaan 95% dari hasil ramalan, sehingga menunjukkan relevansi ramalan dengan data aktual.
=================================================================================================================================
Fluctuations in stock prices are often influenced by various factors, and one of them is sentiment, which can originate from the general public or stock-related news. The stock price of PT. Bank Rakyat Indonesia Tbk. (BBRI), as one of the "big four" largest banks in Indonesia, can also be influenced by this sentiment. In an effort to forecast the closing price of BBRI stock based on news sentiment, this research adopts a machine learning approach using the Support Vector Regression (SVR) method and optimizes the function with the Fruit Fly Optimization Algorithm (FOA). Sentiment is first evaluated using the IndoBERT method. This research proposes several models, including a model without considering sentiment, a model considering sentiment on the day of the event, a model considering sentiment on the previous day, and a model considering sentiment on both the day of the event and the previous day. Sentiment analysis results using IndoBERT show an overall accuracy rate above 90%. Furthermore, modelling results indicate that the best-performing SVR model is the one that does not consider sentiment, with a lower Mean Average Percentage Error (MAPE) compared to other models. The forecasting results from the best model show that the actual data still falls within the 95% confidence interval of the forecast, demonstrating the relevance of the forecast to the actual data.
| Item Type: | Thesis (Diploma) |
|---|---|
| Uncontrolled Keywords: | Saham BBRI, Analisis Sentimen, Support Vector Regression (SVR), Fruit Fly Optimization Algorithm (FOA). |
| Subjects: | T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models. |
| Divisions: | Faculty of Vocational > 49501-Business Statistics |
| Depositing User: | Ferdyansyah Permana Putra |
| Date Deposited: | 01 Feb 2024 07:47 |
| Last Modified: | 01 Feb 2024 07:47 |
| URI: | http://repository.its.ac.id/id/eprint/105901 |
Actions (login required)
 |
View Item |