Scholarlink

<p>The paper applies Convolutional Neural Networks to examine whether and to what extent closing stock prices can be predicted during the opening hour of a trading day. In particular, the <i>MobileNet-V2</i> architecture was implemented, which transforms the financial time series into an image classification problem. We used daily data in a 5-minute time interval of the 1000 largest listings in Nasdaq by market capitalization. Results show that according to a standard performance measures, the <i>MobileNet-V2</i> achieved a high prediction accuracy and outperformed several alternative deep learning algorithms.</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

Using CNN to Model Stock Prices

  • 期刊论文

  • 发表日期：

    2025年2月17日

查看全文
  • Mitja Steinbacher,
  • Matej Steinbacher,
  • Matjaz Steinbacher

摘要

The paper applies Convolutional Neural Networks to examine whether and to what extent closing stock prices can be predicted during the opening hour of a trading day. In particular, the MobileNet-V2 architecture was implemented, which transforms the financial time series into an image classification problem. We used daily data in a 5-minute time interval of the 1000 largest listings in Nasdaq by market capitalization. Results show that according to a standard performance measures, the MobileNet-V2 achieved a high prediction accuracy and outperformed several alternative deep learning algorithms.