摘要
在信息化教育背景下,通过多维度数据融合与深度学习方法,构建高校教学质量分析“数据预处理-特征建模-关联挖掘”的技术链条,实现教学质量分析系统的精准化与智能化。基于DeepSeek-R1模型实现学生评教评语的情感自动检测,有效识别并剔除异常评教数据。运用K-means聚类分析与多重对应分析等多元统计法,挖掘多维评教数据的特征,揭示影响教学质量的关键因素。同时,搭建YOLO系列目标检测深度学习模型,实现学生到课率、抬头率、前排就座率等课堂行为数据的自动化采集,将客观行为数据纳入评价体系,弥补传统主观评价的不足。有效解决了异常数据干扰、分析维度单一、时效滞后等问题,为构建智能化教学质量分析系统提供可靠路径,为改进教学质量提供智能化决策支持。
Under the background of informationized education,a technical chain of“data preprocessing-feature modeling-association mining”for the analysis of teaching quality in colleges and universities is constructed through multi-dimensional data fusion and deep learning methods,achieving the precision and intelligence of the teaching quality analysis system.Based on the DeepSeek-R1 model,the sentiment of students’teaching evaluation comments is automatically detected to effectively identify and eliminate abnormal teaching evaluation data.By using multivariate statistical methods such as K-means clustering analysis and multiple correspondence analysis,the features of multi-dimensional teaching evaluation data are mined to reveal the key factors affecting teaching quality.Meanwhile,a YOLO series of deep learning models for object detection is built to automatically collect classroom behavior data such as student attendance rate,head-up rate,and front-row seating rate,incorporating objective behavior data into the evaluation system to make up for the shortcomings of traditional subjective evaluations.This effectively solves problems such as interference from abnormal data,single analysis dimension,and lagging timeliness,providing a reliable path for building an intelligent teaching quality analysis system and intelligent decision support for improving teaching quality.
作者
汤鸣
杨楠
徐行
TANG Ming;YANG Nan;XU Xing(Academic Affairs Office,Xihang University,Xi′an 710077,China;School of Computer Science,Xihang University,Xi′an 710077,China)
出处
《西安航空学院学报》
2025年第2期56-66,共11页
Journal of Xi’an Aeronautical Institute
基金
西安航空学院高等教育教学改革研究项目(重点项目)(23JXGG1020)。
关键词
多元统计分析
聚类分析
教学质量评价
可视化分析
multivariate statistical analysis
clustering analysis
teaching quality evaluation
visual analysis
