文献类型：期刊文献

英文题名：Deep prediction of marine cultural and creative products purchase intentions by integrating visual significance and textual emotion

作者：Wu, Yifei[1]; Li, Zijun[1]

机构：[1] Zhongge Art College, Guangdong Ocean University, Zhanjiang, 524000, China

年份：2026

卷号：27

期号：54

起止页码：24

外文期刊名：International Journal of Information and Communication Technology

收录：EI(收录号：20262120773072)、Scopus(收录号：2-s2.0-105039750422)

语种：英文

外文关键词：Deep learning - Electronic commerce - Forecasting - Image analysis - Information fusion - Learning systems - Purchasing - Sales - Sentiment analysis - Visualization

外文摘要：Precise prediction of the purchase intention for marine cultural and creative products is of vital importance for e-commerce marketing. Addressing the issues of the existing methods that separately analyse images and comments, and the difficulty in capturing cross-modal collaborative effects, this study proposes a dual-stream deep learning model that integrates visual saliency and text sentiment. This model achieves a deeper understanding of user preferences by simultaneously extracting the salient regions of the images and the sentiment tendencies of the comments. Experiments on public datasets show that the purchase intention prediction accuracy of this method reaches 85.6%, significantly outperforming the baseline models that only use images (72.1%) or text (78.3%), with a recall rate increase of over 10 percentage points. This study provides an effective tool for multimodal fusion analysis and personalised recommendations in the marine cultural and creative field. Copyright ? 2026 Inderscience Enterprises Ltd. Copyright ? The Author(s) 2026. Published by Inderscience Publishers Ltd. This is an Open Access Article distributed under the CC BY license. (http://creativecommons.org/licenses/by/4.0/)