Research Paper Graduate 1,623 words

Clustering Cryptocurrencies: Methods, Results & Trading Insights

~9 min read 6 sections Finance

Abstract

This paper investigates whether fundamental similarities among cryptocurrencies are reflected in market metrics by applying density-based clustering algorithms to time series data. The study develops a pipeline using DBSCAN for top-level cluster identification and HDBSCAN for sub-cluster analysis across bullish, bearish, and stable Bitcoin market states. Feature extraction is performed using both standard statistical parameters and the tsfresh library. The findings indicate that traditional fundamental clustering fails to explain cryptocurrency price dynamics, and that clustering methods must account for varying densities, market states, and high dimensionality. The paper concludes that density-based clustering can inform trading strategies, though further refinement of methods and parameters is needed.

Key Takeaways

Introduction: Why cryptocurrency clustering matters for trading
Developing a Clustering Method: DBSCAN selection and feature extraction pipeline
Clustering Inside Top-Level Clusters: HDBSCAN applied to sub-cluster identification
Features Extraction, Market States, and Results: Clustering across bull, bear, and stable markets
Comparing Clustering Approaches: Limitations of comparing DBSCAN and tsfresh methods
Conclusion: Fundamental clustering rejected; strategy implications

✍️ How to write this paper — guide, tools & examples ▾

What makes this paper effective

The paper moves methodically from problem definition to algorithmic selection to results, giving readers a clear logical progression through a technically complex topic.
The justification for choosing DBSCAN over k-means is clearly argued, grounding the methodological choice in the specific research goal of avoiding presupposed cluster shapes.
The numbered lists and structured part divisions make the multi-stage pipeline easy to follow, which is particularly helpful given the technical subject matter.

Key academic technique demonstrated

The paper demonstrates comparative method evaluation: it applies multiple clustering algorithms (DBSCAN, HDBSCAN, tsfresh-based clustering) and explicitly discusses why direct comparison between them is limited. This honest acknowledgment of methodological incompatibility — rather than forcing a comparison — reflects strong academic judgment and adds credibility to the conclusions.

Structure breakdown

The paper opens with a motivation section establishing why cryptocurrency clustering matters for valuation and trading. It then proceeds through four sequential methodological parts: algorithm selection and justification, sub-cluster analysis using HDBSCAN, feature extraction across market states, and a comparison of clustering approaches. The conclusion synthesizes findings and rejects the fundamental clustering hypothesis, closing with directions for further research.

📘 Read the full research paper guide → Generate citations → Build an outline → Draft a literature review → 📚 More Finance examples →

Essay 1,623 words

Introduction

How to value cryptocurrencies has been a major question ever since so many began finding their way to market. As Quintero (2018) points out, "having a clear and unbiased benchmark while evaluating new decentralized projects in the crypto economy" could help to answer the question of valuation. Clustering commonly occurs around token type; thus, one routinely sees the clustering of currency tokens, platform tokens, utility tokens, brand tokens, and security tokens. Yet these are not the only clusters that may appear the more closely one looks at the space. As clustering shows which cryptocurrencies move in tandem at the top of the market cap, it is useful to examine cryptocurrency clustering to see what similarities in movement might tell us.

Are fundamental similarities backed by market metrics? That is the main question to be asked, and an important one, because clusters can be used to formulate trading strategies. However, Quintero (2018) notes that there is more than one cluster in the cryptocurrency space — in fact, there are numerous ones. Identifying them and understanding the relationships among assets is critical to devising a successful trading strategy. Identifying clusters as part of developing a trading strategy for cryptocurrencies could help make the space far more viable for investors and speculators alike. "There do seem to exist natural clusters of coins that move in tandem," Quintero (2018) states — which means more cryptocurrency samples need to be examined in order to clarify these apparent relationships.

5 Sections Hidden · 1,020 words

Developing a Clustering Method310 words

The problem of time series clustering can be considered as finding a function f(XT) = y ∈ [1…K] for XT = (x1, …, xT) with xT ∈ ℝd, where T is the timeline length and K is the particular cluster. This should be conducted with a representation of the time series…

Clustering Inside Top-Level Clusters200 words

To accomplish clustering inside top-level clusters, which were identified using DBSCAN, Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) was used to perform the DBSCAN algorithm over varying epsilon values and to integrate the results to find a clustering that gives the best stability over epsilon. HDBSCAN allowed the identification of clusters of varying densities (unlike DBSCAN)…

Features Extraction, Market States, and Results210 words

A new instance of the Extractor class can be initiated by passing a periods_of_interest argument, which will slice the scope of each coin along the time axis with respect to periods of interest. This parameter is helpful when executing the bull_bear_periods function. To ascertain…

Comparing Clustering Approaches180 words

Comparing clusters obtained via various methods can be illuminating and offer insights into possible improvements to the created pipeline and directions for further research. There are several reasons why comparing clusters obtained by different methods…

Conclusion120 words

Fundamentals cannot explain cryptocurrency price dynamics. Traditional fundamental clustering does not explain price movement. Therefore, the hypothesis…

View Full Document

References

Quintero, S. (2018). Clustering cryptocurrencies with affinity propagation. Retrieved from

Key Concepts in This Paper

DBSCAN HDBSCAN Time Series Clustering Feature Extraction Market States tsfresh Library Crypto Valuation Outlier Detection Trading Strategy Bitcoin Trends