科技创新与应用2025，Vol.15Issue(35)：49-52,4.DOI:10.19981/j.CN23-1581/G3.2025.35.010
基于深度学习的多模态数据特征提取与选择方法
摘要
Abstract
To solve the problem that traditional machine learning methods cannot effectively utilize unstructured data,there is an urgent need to develop new ways of feature extraction for multimodal data.Based on this,this paper combines and applies the idea of deep learning to design a neural network with the function of multi-modal feature extraction,in order to achieve the effective transformation of heterogeneous modal data to homogeneous modal data,and thereby effectively extract the modal fusion features of the data.Meanwhile,the sparse representation learning algorithm is integrated into the deep neural network to obtain the importance weights of different modalities and screen the features closely related to the current task,so as to eliminate redundant information and noisy data.Finally,the effectiveness of the multi-memory data feature extraction and selection method based on deep learning is verified through experimental analysis.The results show that this method performs well in multimodal feature extraction,modal correlation evaluation,and filtering of redundant and noise information.关键词
深度学习/多模态数据/特征提取/特征选择/神经网络Key words
deep learning/multimodal data/feature extraction/feature selection/neural network分类
信息技术与安全科学引用本文复制引用
..基于深度学习的多模态数据特征提取与选择方法[J].科技创新与应用,2025,15(35):49-52,4.基金项目
2025年江苏省高等教育教改研究课题(2025JGYB795) （2025JGYB795）
2024年盐城市基础研究计划(自然科学基金)面上项目课题(YCBK2024062)阶段成果 （自然科学基金）
