Scholarlink

<p>This paper addresses the challenge of using machine learning to predict stock market trends, by decomposing market behavior using distinct timeframe observational scopes, identifying long-term and short-term trends. Specialized models are developed for each timeframe and are integrated using a dynamic ensemble approach that leverages stock market data. This approach significantly outperforms both traditional machine learning models and the benchmark buy-and-hold strategy, achieving a 258.94% return on investment compared to the buy-and-hold’s 132.11% on the S&amp;P 500 index from 2015 to 2024, demonstrating consistent returns and effective risk management, highlighting its robustness in diverse market conditions. Furthermore, innovative data labeling and hyperparameter tuning methods were developed and shown to be instrumental in improving the performance of the system.</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

Dynamic Ensemble of Specialized Models for Multi-Timeframe Stock Market Trend Prediction

  • 期刊论文

  • 发表日期：

    2026年1月3日

查看全文
  • João Caldeira,
  • Rui Neves

摘要

This paper addresses the challenge of using machine learning to predict stock market trends, by decomposing market behavior using distinct timeframe observational scopes, identifying long-term and short-term trends. Specialized models are developed for each timeframe and are integrated using a dynamic ensemble approach that leverages stock market data. This approach significantly outperforms both traditional machine learning models and the benchmark buy-and-hold strategy, achieving a 258.94% return on investment compared to the buy-and-hold’s 132.11% on the S&P 500 index from 2015 to 2024, demonstrating consistent returns and effective risk management, highlighting its robustness in diverse market conditions. Furthermore, innovative data labeling and hyperparameter tuning methods were developed and shown to be instrumental in improving the performance of the system.