Optimal double deep q-learning network and support vector machine for maximum profit of cryptocurrency trading bot

Hong, Liang Cai (2022) Optimal double deep q-learning network and support vector machine for maximum profit of cryptocurrency trading bot. Masters thesis, Universitas Pelita Harapan.

[img]
Preview
Text (Title)
Title.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (101kB) | Preview
[img]
Preview
Text (Abstract)
Abstract.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (1MB) | Preview
[img]
Preview
Text (ToC)
ToC.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (1MB) | Preview
[img]
Preview
Text (Chapter1)
Chapter1.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (4MB) | Preview
[img] Text (Chapter2)
Chapter2.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (9MB)
[img] Text (Chapter3)
Chapter3.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (7MB)
[img] Text (Chapter4)
Chapter4.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (6MB)
[img] Text (Chapter5)
Chapter5.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (918kB)
[img]
Preview
Text (Bibliography)
Bibliography.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (2MB) | Preview
[img] Text (Appendices)
Binder1_watermark.pdf
Restricted to Repository staff only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (12MB)

Abstract

Since a couple years back, a virtual currency cryptocurrency has grown significantly in popularity. Many parties including investors and scholars alike have developed an interest in this trending subject due to its immense profit potential. Cryptocurrency trading requires a fast yet proper analysis and decision making that relies on the trader to analyze a large amount of data. Many have tried to automate the cryptocurrency trading flow by utilizing various algorithms that fall in the category of prediction models and reinforcement learning to simplify it. In this thesis an automated cryptocurrency trading bot that combines Double Deep Q�Learning Network and Support Vector Machine algorithm to automatically trade effectively in unpredictable situation and gain a sizeable profit is implemented. The trading bot using Double Deep Q-Learning Network and Support Vector Machine produced better profitability. Using hyperparameters of gamma 100 and C 100 in BTC/IDR cryptocurrency pair, gamma 100 and C 100 in ETH/IDR pair, and gamma 80 and C 18 in USDT/IDR pair, the hybrid trading bot scores 24 times more profits in BTC/IDR pair, 52 times more profits in ETH/IDR pair, and 5 times more profit in USD/IDR pair. / Sejak beberapa tahun yang lalu, cryptocurrency mata uang virtual telah tumbuh secara signifikan dalam popularitas. Banyak pihak termasuk investor dan cendekiawan sama-sama tertarik pada topik yang sedang tren ini karena potensi keuntungannya yang sangat besar. Perdagangan cryptocurrency membutuhkan analisis dan pengambilan keputusan yang cepat namun tepat yang bergantung pada pedagang untuk menganalisis sejumlah besar data. Banyak yang mencoba mengotomatiskan alur perdagangan cryptocurrency dengan memanfaatkan berbagai algoritma yang termasuk dalam kategori model prediksi dan pembelajaran penguatan untuk menyederhanakannya. Tesis ini bertujuan untuk membangun bot perdagangan cryptocurrency otomatis yang menggabungkan algoritma Double Deep Q-Learning Network dan Support Vector Machine dengan tujuan untuk berdagang secara otomatis secara efektif dalam situasi yang tidak terduga dan mendapatkan keuntungan yang cukup besar. Bot perdagangan menggunakan Double Deep Q-Learning Network dan Support Vector Machine menghasilkan profitabilitas yang lebih baik. Dengan menggunakan hyperparameter gamma 100 dan C 100 dalam pasangan mata uang kripto BTC/IDR, gamma 100 dan C 100 dalam pasangan ETH/IDR, dan gamma 80 dan C 18 dalam pasangan USDT/IDR, bot perdagangan hibrida mencetak keuntungan 24 kali lebih banyak dalam BTC/ Pasangan IDR, 52 kali lebih banyak keuntungan dalam pasangan ETH/IDR, dan 5 kali lebih banyak keuntungan dalam pasangan USD/IDR.

Item Type: Thesis (Masters)
Creators:
CreatorsNIMEmail
Hong, Liang CaiNIM01679210006liangcai.stdnt@gmail.com
Contributors:
ContributionContributorsNIDN/NIDKEmail
Thesis advisorYugopuspito, PujiantoNIDN0324086701yugopuspito@uph.edu
Thesis advisorTjahyadi, HendraNIDN0410076901hendra.tjahyadi@uph.edu
Uncontrolled Keywords: Cryptocurrency ; Reinformcent Learning ; Support Vector Machine ; Trading
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: University Subject > Current > Faculty/School - UPH Karawaci > School of Information Science and Technology > Master of Informatics
Current > Faculty/School - UPH Karawaci > School of Information Science and Technology > Master of Informatics
Depositing User: Users 29052 not found.
Date Deposited: 16 Feb 2023 00:28
Last Modified: 16 Feb 2023 00:28
URI: http://repository.uph.edu/id/eprint/54420

Actions (login required)

View Item View Item