Malik, Gamar Ramadhani (2026) BITCOIN PRICE MOVEMENT PREDICTION BASED ON ON-CHAIN DATA USING THE BILSTM METHOD. Undergraduate thesis, UPN "Veteran" Jawa Timur.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
22081010229-cover.pdf Download (2MB) |
|
Text (Bab 1)
22081010229-bab1.pdf Download (134kB) |
|
Text (Bab 2)
22081010229-bab2.pdf Restricted to Repository staff only until 15 June 2028. Download (326kB) |
|
Text (Bab 3)
22081010229-bab3.pdf Restricted to Repository staff only until 15 June 2028. Download (512kB) |
|
Text (Bab 4)
22081010229-bab4.pdf Restricted to Repository staff only until 15 June 2028. Download (1MB) |
|
Text (Bab 5)
22081010229-bab5.pdf Download (93kB) |
|
Text (Daftar Pustaka)
22081010229-daftarpustaka.pdf Download (155kB) |
|
Text (Lampiran)
22081010229-lampiran.pdf Restricted to Repository staff only until 15 June 2028. Download (666kB) |
Abstract
Bitcoin is a decentralized cryptocurrency built on Blockchain technology that exhibits extremely high price volatility, making it difficult for investors to make sound investment decisions. This study aims to develop a Bitcoin price prediction model using the Bidirectional Long Short-Term Memory (BiLSTM) method based on on chain data, including transaction volume, tx_count, active_spending_addresses, and fee_btc. The BiLSTM approach was chosen for its ability to process time series data bidirectionally, making it more effective at capturing extreme volatility patterns compared to conventional LSTM. The dataset was obtained from BigQuery and underwent preprocessing stages including data transformation, normalization, data splitting, and sequence formation. Experiments were conducted by varying the data split ratio, learning rate, and number of hidden layers. The optimal model was achieved using a 2-hidden-layer architecture, a learning rate of 0.0001, and a 90:10 data split, yielding MAE of 932.61, RMSE of 1,257.07, and MAPE of 0.97% classified as a highly accurate prediction. The model was subsequently deployed as a Flask-based web application in the form of an interactive single-page dashboard, enabling users to upload datasets, execute price forecasts, and visualize market analysis results in a unified interface. Keyword: Bitcoin, BiLSTM, Data On-Chain, Forecasting, Deep Learning
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Gamar Ramadhani Malik | ||||||||||||
| Date Deposited: | 15 Jun 2026 06:24 | ||||||||||||
| Last Modified: | 15 Jun 2026 06:24 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/54003 |
Actions (login required)
 |
View Item |