Scholarlink

<p>We introduce a new deep-learning method to perform jump detection on financial time series data. We train a convolution neural network (CNN) model with simulated data based on a mixture of various stochastic models. This method can equip the CNN model with great generalization ability for out-of-sample jump detection. Experiments on simulated and real data show that the proposed method strongly outperforms the traditional statistic test method, achieving higher detection accuracy.</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

Jump Detection Using Deep Learning: With Applications to Financial Time Series Data

  • 期刊论文

  • 发表日期：

    2025年4月29日

查看全文
  • Weizheng Chen,
  • Guang Zhang

摘要

We introduce a new deep-learning method to perform jump detection on financial time series data. We train a convolution neural network (CNN) model with simulated data based on a mixture of various stochastic models. This method can equip the CNN model with great generalization ability for out-of-sample jump detection. Experiments on simulated and real data show that the proposed method strongly outperforms the traditional statistic test method, achieving higher detection accuracy.