多模态数据与机器学习赋能高职教师数字画像构建  

Multimodal data and machine learning empowering the construction of digital profiles for vocational college teacher

作　　者：黄冠森 满硕 HUANG Guansen;MAN Shuo(Shijiazhuang Posts and Telecommunications Technical College,Shijiazhuang 050022,China)

机构地区：[1]石家庄邮电职业技术学院,石家庄050022

出　　处：《计算机应用文摘》2025年第9期120-122,共3页

摘　　要：文章提出一种基于多模态数据融合的教师数字画像系统。它包含三阶段处理流程,首先通过多层次采集机制和增量同步策略实现多源教师数据的实时精准采集;其次利用改进的CILP框架完成跨模态数据的特征提取与降维,有效解决多模态数据对齐与融合问题;最后结合IQR数据清洗和自监督深度聚类算法生成多维度的教师精准画像。实验结果表明,该系统能够有效整合教师教学、科研、管理等多维度数据,为高职院校教师评价与发展提供数据支持。This article proposes a teacher digital portrait system based on multimodal data fusion.It includes a three-stage processing flow,which first achieves real-time and accurate collection of multi-source teacher data through a multi-level collection mechanism and incremental synchronization strategy.Secondly,the improved CILP framework is utilized to achieve feature extraction and dimensionality reduction of cross modal data,effectively solving the problem of alignment and fusion of multimodal data.Finally,combining IQR data cleaning and self supervised deep clustering algorithm to generate multi-dimensional accurate teacher portraits.The experimental results show that the system can effectively integrate multi-dimensional data such as teacher teaching,research,and management,providing data support for the evaluation and development of vocational college teachers.

关 键 词：机器学习 教师画像 多模态数据 高职院校 数据分析 

分 类 号：G717[文化科学—职业技术教育学]