文献类型：期刊文献

英文题名：Digital museum intelligent interactive system based on SSA-transformer-GRU algorithm☆

作者：Mao, Kainan Zou, Yutian Xiang, Kun Huang, Yongling

第一作者：毛凯楠;Mao, Kainan

通信作者：Huang, YL[1]

机构：[1]Guizhou Inst Technol, Sch Resources & Environm Engn, Guiyang 550025, Peoples R China;[2]Guiyang Vocat & Tech Coll, Guiyang 550081, Peoples R China;[3]Geol Museum Guizhou, Guiyang 550081, Peoples R China

第一机构：贵州理工学院资源与环境工程学院

通信机构：corresponding author), Geol Museum Guizhou, Guiyang 550081, Peoples R China.

年份：2026

卷号：58

外文期刊名：ENTERTAINMENT COMPUTING

收录：;EI(收录号：20263021144728);Scopus(收录号：2-s2.0-105044893028);WOS:【SCI-EXPANDED(收录号:WOS:001830616200001)】；

基金：This work was supported by Guizhou Provincial Key Technology R & D Program [2023] GEN 167 (No.) ; Guizhou Provincial Key Tech-nology R & D Program [2023] GEN 168 (No.) ; Guizhou Institute of Technology High-Level Talent Scientific Research Start-up Fund Project (XJGC20190946) .

语种：英文

外文关键词：Sparrow search algorithm; Transformer; GRU; Digital museum; Intelligent interaction; Multimodal data; Gated recurrent unit

摘要：In the current development of digital museums, intelligent interactive systems face the problems of low multimodal data processing efficiency and insufficient user response accuracy. In this paper, an intelligent interactive system based on the SSA-Transformer-GRU algorithm is constructed. By integrating the multi-modal feature extraction capabilities of Transformer, the sequence modeling advantages of GRU, and the parameter optimization mechanism of SSA, the system performance is significantly improved. There are three innovations in this study: firstly, the SSA-Transformer-GRU hybrid algorithm architecture is proposed to realize the automatic optimization of model parameters; Secondly, a multi-modal data fusion processing process is constructed to solve the problem of heterogeneity of cultural relics data; Finally, the intelligent interactive system framework is designed, which integrates knowledge graph and recommendation functions. The experiment uses the real interactive data set of a digital museum, and the system demonstrated a response accuracy of 94.2% and an F1 score of 92.8%, the response time is 1.3 s, and the user experience satisfaction score is 4.6/5.0; all indicators are better than the traditional method. This system effectively improves the interactive accuracy and user experience of digital museums and provides technical support for the digitization of cultural heritage. In the future, the application of algorithms in more scenarios will be explored.