Predicting Price Trends Using Sentiment Analysis: A Study of StepN’s SocialFi and GameFi Cryptocurrencies




cryptocurrency, sentiment analysis, machine learning, fintech, investment


The cryptocurrency market, specifically the non-fungible token (NFT) market, has been gaining popularity with the rise of social finance, game finance, metaverse, and web 3.0 technologies. With the increasing interest in cryptocurrency, it is essential to develop a comprehensive understanding of the market dynamics to aid investment decisions. This paper aims to analyze the impact of news sentiment on the prices of two cryptocurrencies, Green Satoshi Token (GST) and Green Metaverse Token (GMT). The sentiment analysis model used in this study is Finance Bidirectional Encoder Representations from Transformers (FinBERT), a pre-trained deep neural network model designed for financial sentiment analysis. Additionally, we introduce the use of the Extreme Gradient Boosting (XGBoost) algorithm to evaluate the sentiment result on the model’s performance. The study period covered from March 2022 to April 2022, and the sentiment score of the result generated by FinBERT on crypto, stock market, and finance news was found to be correlated with the prices of GST and GMT. The findings suggest that the sentiment score of GST reflects changes in the price earlier than GMT. These findings have significant implications for decision-making strategies and can aid investors in making more informed decisions. The research highlights the importance of sentiment analysis in understanding the market dynamics and its potential impact on the prices of cryptocurrencies. The use of FinBERT and XGBoost algorithms provides valuable insights into market trends and can aid investors in making informed decisions.

Author Biography

Tinfah Chung, HELP University, Kuala Lumpur, Malaysia







How to Cite

Yeoh ED, Chung T, Wang Y. Predicting Price Trends Using Sentiment Analysis: A Study of StepN’s SocialFi and GameFi Cryptocurrencies. Contemp. Math. [Internet]. 2023 Nov. 15 [cited 2024 May 28];4(4):1089-108. Available from: