Figure Ilustration AI
FORMOSA NEWS - Kazakhstan - Multilingual Sentiment Index Reveals Hidden Signals in Bitcoin and Ethereum Price Movements. A 2026 study by Ningyu Zhou of Al-Farabi Kazakh National University, Kazakhstan, shows that global news sentiment across multiple languages can influence short-term price movements of Bitcoin (BTC) and Ethereum (ETH). Published in the Journal of Finance and Business Digital (JFBD), the research introduces a new tool the Multilingual Crypto Sentiment Index (ML-CryptoSI) that captures investor sentiment from international news sources. The findings matter because cryptocurrency markets are highly sensitive to narratives and public perception, often reacting faster to sentiment than to traditional financial indicators.
Global Narratives Shape Crypto Markets
Cryptocurrency markets have experienced dramatic cycles of boom and crash over the past decade. Unlike traditional financial assets, Bitcoin and Ethereum lack stable income streams or widely accepted valuation benchmarks. This makes them especially vulnerable to shifts in investor mood, media narratives, and public attention. Previous studies have linked sentiment especially from English-language media and social platforms to crypto price movements. However, these approaches often overlook the global nature of cryptocurrency markets, where information flows in multiple languages and across regions. Ningyu Zhou addresses this gap by integrating multilingual news data into a unified sentiment index.
How the Study Was Conducted
The research combines large-scale text analysis with financial market data. Zhou collected more than 22,000 cryptocurrency-related news articles published in 2025 from major global media outlets, including CoinDesk, Cointelegraph, Odaily, BlockBeats, and PANews. After filtering for quality and language consistency, about 18,700 articles in six languages were retained: English, Chinese, Korean, Japanese, Spanish, and Russian.
Each article was analyzed using advanced language models to classify sentiment as positive, negative, or neutral. The results show a strong dominance of neutral reporting:
- Neutral sentiment: 83.6%.
- Negative sentiment: 10.2%.
- Positive sentiment: 6.2%.
These individual sentiment scores were then aggregated into a daily index. Zhou applied a statistical method called Principal Component Analysis (PCA) to extract a common sentiment factor across all languages, resulting in the ML-CryptoSI. The study also incorporated daily price, volatility, and trading volume data for Bitcoin and Ethereum from Binance and CoinGecko. By aligning sentiment data with market data, the research tested whether sentiment could predict next-day returns and risk levels.
Key Findings: Stronger Impact on Ethereum
The analysis reveals that multilingual sentiment carries measurable predictive power particularly for Ethereum.
Key findings include:
- ML-CryptoSI predicts next-day returns for Ethereum, with a statistically meaningful relationship.
- The predictive effect is weaker and not statistically significant for Bitcoin.
- Sentiment impact increases during periods of high news activity, suggesting that information intensity amplifies market reactions.
- No strong evidence links sentiment to volatility, indicating that price fluctuations may depend on other structural factors.
Interestingly, the relationship between sentiment and price is negative in the short term. When sentiment becomes more positive, prices tend to decline slightly the following day. This pattern suggests short-term overreaction, where markets initially respond strongly to news and then correct.
Real-World Implications for Investors and Industry
The introduction of ML-CryptoSI has practical implications for cryptocurrency markets and financial technology:
- Smarter Trading Strategies - Investors can use sentiment indicators as short-term signals, especially during periods of intense news coverage.
- Enhanced Risk Management - Understanding sentiment-driven price reversals can help traders anticipate corrections and reduce losses.
- Focus on Ethereum Dynamics - The stronger sensitivity of Ethereum to sentiment suggests it may offer more opportunities for sentiment-based trading strategies.
- AI-Driven Financial Models - The study demonstrates how artificial intelligence and multilingual data can improve forecasting tools in digital finance.
Author Profile
Ningyu Zhou is a researcher at Al-Farabi Kazakh National University, Kazakhstan. Zhou specializes in digital finance, sentiment analysis, and the application of artificial intelligence in financial markets, with a focus on cryptocurrency pricing and risk modeling.
Source
Zhou, N. (2026). ML-CryptoSI: A Multilingual Crypto Sentiment Index and its Role in Bitcoin and Ethereum Pricing. Journal of Finance and Business Digital (JFBD), Vol. 5 No. 1, pp. 33–52.
DOI: https://doi.org/10.55927/jfbd.v5i1.6
URL: https://journaljfbd.my.id/index.php/jfbd
0 Komentar