Analisis Clustering Pada Harga Koin Cryptocurrency Dengan Jarak Dynamic Time Warping

Nabilah, Jihan (2023) Analisis Clustering Pada Harga Koin Cryptocurrency Dengan Jarak Dynamic Time Warping. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211940000068-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (2MB) | Request a copy

Abstract

Pertumbuhan perekonomian dunia yang diprediksi akan memburuk pada tahun 2023 menyebabkan kemungkinan dunia mengalami resesi global. Resesi ini dipicu oleh keadaan bank sentral negara yang secara bersamaan menaikkan suku bunga sebagai respons inflasi. Berdasarkan survey Bloomberg, Indonesia memiliki risiko resesi hanya sebesar 3%. Namun, tidak menutup kemungkinan bahwa Indonesia dapat lepas dari guncangan resesi dunia. Investasi adalah salah satu jalan bagi masyarakat ketika dihadapkan pada peristiwa resesi. Salah satu investasi dengan prospek baik masa depan adalah investasi cryptocurrency. Walaupun demikian, untuk meminimalisir adanya kerugian yang dapat dialami perlu melihat perkembangan cyptocurrency pada beberapa tahun terakhir. Salah satu cara untuk dapat melihat perkembangan cryptocurrency adalah dengan melakukan pengelompokan atau clustering cryptocurrency. Metode yang sesuai untuk data deret waktu cryptocurrency yang digunakan adalah jarak Dynamic Time Warping (DTW). Oleh karena itu, dilakukan penelitian clustering time series harga cryptocurrency dengan menggunakan jarak DTW. Klaster terbaik dipilih berdasarkan koefisien silhouette terbesar. Pada penelitian ini didapatkan klaster optimum terbaik dengan menggunakan metode clustering hirarki average linkage yang menghasilkan koefisien silhouette sebesar 0.486. Pola pergerakan harga cryptocurrency terbagi menjadi 2 klaster yakni klaster pertama yang merupakan klaster dengan pola historical harga naik dan kluster kedua yang memiliki pola historical harga menurun.
=================================================================================================================================
World economic growth is predicted to worsen in 2023, leading to the possibility of a global recession. This recession is triggered by the state of the country's central banks which simultaneously raised interest rates in response to inflation. Bloomberg survey shows that Indonesia has a recession risk of only 3%, but it does not rule out the possibility that Indonesia can escape the effects of a world recession. Investment is considered one of the ways for people facing global recession. One example of investment with good prospects for the future is cryptocurrency investment. However, to minimize the losses that people can experience, it is necessary to look at the development of cryptocurrency in the past few years. One way to see the development of cryptocurrencies is by grouping or clustering cryptocurrencies. A suitable distance for cryptocurrency time series data is Dynamic Time Warping (DTW) distance. Therefore, research on clustering time series cryptocurrency prices was carried out using DTW distances. The optimum cluster is chosen based on the largest silhouette coefficient. In this research, the best optimum cluster is obtained using the average linkage hierarchical clustering method which produces a silhouette coefficient of 0.486. The price movement pattern of cryptocurrencies is divided into 2 clusters, which are the first cluster with a historical price growth pattern and the second cluster with a historical price decline pattern.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analisis Clustering, Cryptocurrency, Dynamic Time Warping, Silhouette, Clustering Analysis
Subjects:	Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Jihan Nabilah
Date Deposited:	25 Sep 2023 07:56
Last Modified:	25 Sep 2023 07:56
URI:	http://repository.its.ac.id/id/eprint/104558

Actions (login required)

View Item