Scholarlink

<p>The cryptocurrency market’s high volatility and unpredictable price movements create significant challenges for traders. Traditional financial models often fail to capture these dynamics, leading to suboptimal investment strategies. The purpose of this study is to introduce YOLOFin, a financial trading optimization model based on the YOLOv8 deep learning architecture, marking its first application in cryptocurrency trading. YOLOFin transforms financial time-series data into visual representations using five different image types (Bar Chart, Candlestick, Gramian Angular Field, Heatmap, and Multi-Chart) to enhance pattern recognition. The model employs innovative image processing and labeling strategies to capture both short- and long-term price trends, enabling adaptive investment strategies. YOLOFin analyzes the past 30 days of price, volume, and technical indicator data to estimate the average return for the next 15 days, generating trading signals for Buy, Sell, or Hold decisions. By leveraging YOLO-based image classification, the model effectively addresses the challenges posed by high volatility and unpredictable price movements while achieving 84–87% classification accuracy and delivering a robust Compound Annual Growth Rate (CAGR), with BNB reaching 276% and BTC 143%. By integrating image-based deep learning with adaptive trading strategies, YOLOFin offers a structured and stable approach to navigating volatile cryptocurrency markets. Future research will explore reinforcement learning and sentiment analysis to enhance predictive capabilities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A new YOLO-based financial trading optimization model: YOLOFin—a case study in the cryptocurrency market

  • 期刊论文

  • 发表日期：

    2025年4月7日

查看全文
  • Muhammet Rıdvan İNCE,
  • Meltem KURT PEHLİVANOĞLU

摘要

The cryptocurrency market’s high volatility and unpredictable price movements create significant challenges for traders. Traditional financial models often fail to capture these dynamics, leading to suboptimal investment strategies. The purpose of this study is to introduce YOLOFin, a financial trading optimization model based on the YOLOv8 deep learning architecture, marking its first application in cryptocurrency trading. YOLOFin transforms financial time-series data into visual representations using five different image types (Bar Chart, Candlestick, Gramian Angular Field, Heatmap, and Multi-Chart) to enhance pattern recognition. The model employs innovative image processing and labeling strategies to capture both short- and long-term price trends, enabling adaptive investment strategies. YOLOFin analyzes the past 30 days of price, volume, and technical indicator data to estimate the average return for the next 15 days, generating trading signals for Buy, Sell, or Hold decisions. By leveraging YOLO-based image classification, the model effectively addresses the challenges posed by high volatility and unpredictable price movements while achieving 84–87% classification accuracy and delivering a robust Compound Annual Growth Rate (CAGR), with BNB reaching 276% and BTC 143%. By integrating image-based deep learning with adaptive trading strategies, YOLOFin offers a structured and stable approach to navigating volatile cryptocurrency markets. Future research will explore reinforcement learning and sentiment analysis to enhance predictive capabilities.