FINANCIAL BUBBLE ANALYSIS IN CRYPTOCURRENCY USING LOG PERIODIC POWER LAW SINGULARITY WITH TWO STEPS NONLINEAR OPTIMIZATION AND MINIMUM SPANNING TREE

• Main Author: Alodia Chrisdiana, Kezia
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81521
• Institution: Institut Teknologi Bandung

Econophysics is a physical science that can be used to enrich perspectives on economic problems. One application of econophysics is to analyze the phenomenon of financial bubbles. Financial bubbles are a phenomenon experienced by many assets, including cryptocurrencies. Therefore, this study aims to analyze financial bubbles in cryptocurrencies using the Log Periodic Power Law Singularity (LPPLS) model with a two steps nonlinear optimization fitting method. The research seeks to model and determine the critical time for cryptocurrencies. Additionally, this study analyzes the network topology of 10 sample cryptocurrencies using the Minimum Spanning Tree formed by Prim's algorithm. The study also examines centrality by determining degree centrality, betweenness centrality, and closeness centrality parameters. The sample for this study consists of 10 types of cryptocurrencies: BTC, ETH, ETC, NEO, DASH, XMR, XRP, XLM, GNO, and VTC. The data used in this research include daily historical price data, daily transaction volume data, and daily tweet volume data from January 1, 2019, to April 30, 2021. Based on the LPPLS model fitting results, a financial bubble pattern was identified in the 10 samples, and the critical time predictions had an error range of 1 to 27 days. Additionally, it was found that BTC and ETH have the highest centrality, with a degree centrality value of 0.333.