Putri, Diva Ramadhani Ristiaji (2026) A DECISION SUPPORT SYSTEM FOR BITCOIN PRICE PREDICTION BASED ON THE COMBINATION OF PATTERNED DATASETS, BTC DOMINANCE, AND THE FEAR & GREED INDEX. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (cover)
22081010227.-cover.pdf Download (1MB) | Preview |
|
|
Text (Bab 1)
22081010227.-bab1.pdf Download (167kB) | Preview |
|
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Bab 2)
22081010227.-bab2.pdf Restricted to Repository staff only until 31 May 2028. Download (526kB) |
|
|
Text (Bab 3)
22081010227.-bab3.pdf Restricted to Repository staff only until 31 May 2028. Download (1MB) |
|
|
Text (Bab 4)
22081010227.-bab4.pdf Restricted to Repository staff only until 31 May 2028. Download (1MB) |
|
|
Text (Bab 5)
22081010227.-bab5.pdf Download (174kB) | Preview |
|
|
Text (Daftar Pustaka)
22081010227.-daftarpustaka.pdf Download (180kB) | Preview |
|
|
Text (Lampiran)
22081010227.-lampiran.pdf Restricted to Repository staff only until 31 May 2028. Download (390kB) |
Abstract
High Bitcoin price volatility and the complexity of market factors make it difficult to interpret price movement directions using only a single indicator. Cryptocurrency market analysis is influenced by technical aspects, market structure, and investor sentiment, thus requiring a multi-indicator approach that is more structured and easier to interpret. However, most previous studies still focus on the separate use of indicators or on complex and less interpretable machine learning approaches. This study develops a web-based Decision Support System (DSS) that integrates patterned dataset, Bitcoin Dominance (BTC. D), and Fear & Greed Index (FGI). The dataset used consists of 1-hour Bitcoin OHLCV data from January 1st, 2022 to November 16th, 2025. The patterned dataset is constructed using Range (R), Top Range (TR), Lower Range (LR), as well as PTR and PLR ratios to identify Diamond Crash and Diamond Moon patterns. BTC. D is classified using a quantile-based approach into low, medium, and high categories, while FGI is grouped into five sentiment categories. The indicator combinations are analyzed using a rule-based approach and extracted into IF–THEN rules as the core of the system, supported by the ARIMAX model for short-term price estimation. The results show that the multi-aspect indicator combination is capable of generating highly reliable market direction signals. The rule-based system achieved an accuracy of 97.27%, an average ROI of 8.71%, a win rate of 97.27%, and a coverage of 100%. Further testing on new data from November 2025 to April 2026 also demonstrated consistent performance with an accuracy ranging from 95.65% to 97.55%. The developed system is capable of presenting market condition information, rule-based signals, indicator visualizations, and price estimations through a single web-based interface to help users understand cryptocurrency market conditions in a more structured manner.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > QA Mathematics > QA76 Computer software | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Diva Ramadhani Ristiaji | ||||||||||||
| Date Deposited: | 02 Jun 2026 08:03 | ||||||||||||
| Last Modified: | 02 Jun 2026 08:52 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/53323 |
Actions (login required)
 |
View Item |