Enhancing Cryptocurrency Sentiment Analysis with Multimodal Features

2025-08-17

AI Summary

Existing cryptocurrency sentiment analysis predominantly focuses on textual platforms (e.g., Twitter), overlooking rich affective and contextual signals embedded in video platforms such as TikTok—leading to incomplete market sentiment characterization. This paper presents the first systematic comparative study of video versus text modalities for cryptocurrency sentiment prediction. We propose a novel multimodal sentiment analysis framework that jointly leverages TikTok videos and Twitter texts: cross-modal features are extracted via large language models and vision encoders, then integrated through temporal modeling and cross-platform sentiment fusion to enable fine-grained sentiment tracking and short-term market trend forecasting. Empirical results demonstrate that TikTok-derived sentiment signals exhibit significant explanatory power for intraday price volatility. Our model achieves up to a 20% improvement in prediction accuracy over unimodal baselines, substantiating the unique and indispensable value of video content in financial sentiment analysis.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Economics, Online Markets and Human Computation: Economic aspects of blockchain and cryptocurrencies
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

As cryptocurrencies gain popularity, the digital asset marketplace becomes increasingly significant. Understanding social media signals offers valuable insights into investor sentiment and market dynamics. Prior research has predominantly focused on text-based platforms such as Twitter. However, video content remains underexplored, despite potentially containing richer emotional and contextual sentiment that is not fully captured by text alone. In this study, we present a multimodal analysis comparing TikTok and Twitter sentiment, using large language models to extract insights from both video and text data. We investigate the dynamic dependencies and spillover effects between social media sentiment and cryptocurrency market indicators. Our results reveal that TikTok's video-based sentiment significantly influences speculative assets and short-term market trends, while Twitter's text-based sentiment aligns more closely with long-term dynamics. Notably, the integration of cross-platform sentiment signals improves forecasting accuracy by up to 20%.

Problem

• Analyzing multimodal sentiment from TikTok and Twitter
• Investigating sentiment's impact on cryptocurrency market dynamics
• Improving forecasting accuracy with cross-platform sentiment integration

Innovation

• Multimodal analysis comparing TikTok and Twitter sentiment
• Using large language models for video and text data
• Integration of cross-platform sentiment signals improves forecasting