Paper: arXiv 2308.08554
Authors: Abdulrezzak Zekiye, Semih Utku, Fadi Amroush, Oznur Ozkasap
Abstract
Cryptocurrencies have become a popular and widely researched topic of interest in recent years for investors and scholars. In order to make informed investment decisions, it is essential to comprehend the factors that impact cryptocurrency prices and to identify risky cryptocurrencies. This paper focuses on analyzing historical data and using artificial intelligence algorithms on on-chain parameters to identify the factors affecting a cryptocurrency’s price and to find risky cryptocurrencies. We conducted an analysis of historical cryptocurrencies’ on-chain data and measured the correlation between the price and other parameters. In addition, we used clustering and classification in order to get a better understanding of a cryptocurrency and classify it as risky or not. The analysis revealed that a significant proportion of cryptocurrencies (39%) disappeared from the market, while only a small fraction (10%) survived for more than 1000 days. Our analysis revealed a significant negative correlation between cryptocurrency price and maximum and total supply, as well as a weak positive correlation between price and 24-hour trading volume. Moreover, we clustered cryptocurrencies into five distinct groups using their on-chain parameters, which provides investors with a more comprehensive understanding of a cryptocurrency when compared to those clustered with it. Finally, by implementing multiple classifiers to predict whether a cryptocurrency is risky or not, we obtained the best f1-score of 76% using K-Nearest Neighbor.
Complexity vs Empirical Score
- Math Complexity: 4.0/10
- Empirical Rigor: 7.5/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper applies standard statistical methods (correlation, K-means, K-NN) rather than advanced mathematics, but is heavily data-driven with specific metrics (F1-score, timeframes, market data sources) and a focus on backtest-ready classification.
Research Flowchart
flowchart TD
A["Research Goal:<br>Identify Price Factors &<br>Risky Cryptocurrencies"] --> B["Data Collection<br>Historical On-Chain Data"]
B --> C{"Computational Analysis"}
C --> D["Correlation Analysis<br>Price vs. Max/Total Supply<br>and 24h Volume"]
C --> E["Clustering<br>K-Means: 5 Distinct Groups"]
C --> F["Classification<br>Classifiers for Risk Prediction"]
D --> G["Key Findings"]
E --> G
F --> G
subgraph G ["Outcomes"]
H["Negative Correlation:<br>Price vs. Supply"]
I["Positive Correlation:<br>Price vs. 24h Volume"]
J["Risk Prediction:<br>76% F1-Score (KNN)"]
K["Market Survival:<br>39% Disappeared, 10% >1000 Days"]
end
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes using AI techniques, including clustering and classification on on‑chain parameters, to identify factors affecting cryptocurrency prices and to classify cryptocurrencies as risky or not.
Source sentence
We used clustering and classification in order to get a better understanding of a cryptocurrency and classify it as risky or not.
Assumes 5
- A cryptocurrency that has disappeared from the market is labeled as risky. data
Source sentence
We labeled the cryptocurrencies that disappeared from the market as risky ones and trained our model based on them.
- The classes (risky and not risky) are imbalanced, so f1‑score is the primary metric. statistical
Source sentence
Among the four metrics, we considered f1-score to choose the best classifier since the classes (risky and not risky) are imbalanced.
- Random splitting of the data (80% train, 20% test) is appropriate for all classifiers except LSTM. execution
Source sentence
Random splitting was employed for all classifiers, except for the Long-Short Term Memory classifier.
- Normalization by dividing each feature by its maximum value is sufficient for clustering. model
Source sentence
The PHP-ML library was used to normalize the features by dividing each one by its maximum value.
Show 1 more
- Default settings of scikit‑learn classifiers are adequate for the classification task. model
Source sentence
Various classifiers with their default settings from the scikit‑learn library were employed in this study.
Data
- Historical on‑chain data crypto-onchain unknown
Source sentence
we had to limit the data processing to historical data spanning from 2013 to 2020.
Costs
Not applicable
Finds 3
- Approximately 39% of cryptocurrencies disappeared from the market, while only about 10% survived more than 1000 days. statistical
Source sentence
The analysis revealed that a significant proportion of cryptocurrencies (39%) disappeared from the market, while only a small fraction (10%) survived for more than 1000 days.
- Price shows a significant negative correlation with maximum and total supply, and a weak positive correlation with 24‑hour trading volume. statistical
Source sentence
Our analysis revealed a significant negative correlation between cryptocurrency price and maximum and total supply, as well as a weak positive correlation between price and 24-hour trading volume.
- The K‑Nearest Neighbor classifier achieved the highest f1‑score of approximately 76% for predicting risky cryptocurrencies. predictive-accuracy
Source sentence
Based on the results, the K-Nearest Neighbour classifier achieved the highest f1-Score at approximately 76%.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 19 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.