Scholarlink

<p>Bitcoin, a decentralized digital currency, has garnered significant attention in the financial world. The extreme volatility of bitcoin prices presents both opportunities and risks for investors. To navigate this landscape, accurate price forecasting and the identification of anomalies are paramount. This research delves into the application of anomaly detection techniques to enhance bitcoin price forecasting, leveraging the strengths of multiple algorithms. The study primarily focuses on the design and implementation of Long Short-Term Memory (LSTM) and Density-based Spatial Clustering of Applications with Noise (DBSCAN) models. To achieve these objectives, historical Bitcoin price data in USD is collected, preprocessed, and used to train and evaluate the LSTM and DBSCAN-based models. For anomaly detection on Bitcoin price, we specifically focused on the LSTM-type model and DBSCAN techniques. Anomaly detection techniques, such as statistical analysis and DBSCAN, are integrated into the models to identify and flag instances of unusual market behavior. The research evaluates the effectiveness of these models in enhancing the accuracy and robustness of price predictions while effectively identifying and handling anomalies. The findings from this research contribute to the growing body of knowledge on cryptocurrency price forecasting by integrating the algorithms with advanced anomaly detection techniques.</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

An optimized anomaly detection for bitcoin prices using LSTM and DBSCAN methods

  • 期刊论文

  • 发表日期：

    2025年8月19日

查看全文
  • Aditya Bhushan,
  • Ashutosh Kumar Singh

摘要

Bitcoin, a decentralized digital currency, has garnered significant attention in the financial world. The extreme volatility of bitcoin prices presents both opportunities and risks for investors. To navigate this landscape, accurate price forecasting and the identification of anomalies are paramount. This research delves into the application of anomaly detection techniques to enhance bitcoin price forecasting, leveraging the strengths of multiple algorithms. The study primarily focuses on the design and implementation of Long Short-Term Memory (LSTM) and Density-based Spatial Clustering of Applications with Noise (DBSCAN) models. To achieve these objectives, historical Bitcoin price data in USD is collected, preprocessed, and used to train and evaluate the LSTM and DBSCAN-based models. For anomaly detection on Bitcoin price, we specifically focused on the LSTM-type model and DBSCAN techniques. Anomaly detection techniques, such as statistical analysis and DBSCAN, are integrated into the models to identify and flag instances of unusual market behavior. The research evaluates the effectiveness of these models in enhancing the accuracy and robustness of price predictions while effectively identifying and handling anomalies. The findings from this research contribute to the growing body of knowledge on cryptocurrency price forecasting by integrating the algorithms with advanced anomaly detection techniques.