Scholarlink

<p>The volatility of cryptocurrency markets has sparked significant interest in developing predictive models capable of accurately forecasting price movements. Addressing the complexities posed by the non-linear and dynamic nature of cryptocurrency price data, this study introduces a hybrid model that combines one-dimensional convolutional neural networks (1D-CNN) and long short-term memory (LSTM) networks to forecast Ethereum’s closing prices. The 1D-CNN component captures localized patterns within the time-series data, while the LSTM component effectively models long-term dependencies and sequential trends. A sliding window technique is applied to preprocess Ethereum trading data, enabling the model to manage temporal structures and enhance predictive accuracy. Experiments were performed to assess the hybrid model’s performance under various configurations, benchmarking it against standalone 1D-CNN and LSTM models. Experimental results demonstrated that the hybrid model significantly outperforms these baseline models, promising its potential for improving forecasting accuracy in cryptocurrency markets.</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

Hybrid model of 1D-CNN and LSTM for forecasting Ethereum closing prices: a case study of temporal analysis

  • 期刊论文

  • 发表日期：

    2025年2月26日

查看全文
  • Tran Thai Hoa,
  • Thanh Manh Le,
  • Cuong H. Nguyen-Dinh

摘要

The volatility of cryptocurrency markets has sparked significant interest in developing predictive models capable of accurately forecasting price movements. Addressing the complexities posed by the non-linear and dynamic nature of cryptocurrency price data, this study introduces a hybrid model that combines one-dimensional convolutional neural networks (1D-CNN) and long short-term memory (LSTM) networks to forecast Ethereum’s closing prices. The 1D-CNN component captures localized patterns within the time-series data, while the LSTM component effectively models long-term dependencies and sequential trends. A sliding window technique is applied to preprocess Ethereum trading data, enabling the model to manage temporal structures and enhance predictive accuracy. Experiments were performed to assess the hybrid model’s performance under various configurations, benchmarking it against standalone 1D-CNN and LSTM models. Experimental results demonstrated that the hybrid model significantly outperforms these baseline models, promising its potential for improving forecasting accuracy in cryptocurrency markets.