Changchun Guanghua University, Changchun 130031, Jilin, China.
Comput Intell Neurosci. 2022 Aug 4;2022:2754626. doi: 10.1155/2022/2754626. eCollection 2022.
The current education evaluation is limited not only to the mode of simplification, indexing, and datafication, but also to the scientific nature of college teaching evaluation. This work firstly conducts a theoretical analysis of natural language processing technology, analyzes the related technologies of intelligent scoring, designs a systematic process for intelligent scoring of college English teaching, and finally conducts theoretical research on the Naive Bayesian algorithm in machine learning. In addition, the error of intelligent scoring of English teaching in colleges and universities and the accuracy of scoring and classification are analyzed and researched. The results show that the error between manual scoring and machine scoring is basically about 2 points and the minimum error of intelligent scoring in college English teaching under machine scoring can reach 0 points. There is a certain bias in manual scoring, and scoring on the machine can reduce the generation of this error. The Naive Bayes algorithm has the highest classification accuracy on the college intelligent scoring dataset, which is 76.43%. The weighted Naive Bayes algorithm has been improved in the classification accuracy of college English teaching intelligent scoring, with an average accuracy rate of 74.87%. To sum up, the weighted Naive Bayes algorithm has better performance in the classification accuracy of college English intelligent scoring. This work has a significant effect on the scoring of the college intelligent teaching scoring system under natural language processing and the classification of college teaching intelligence scoring under the Naive Bayes algorithm, which can improve the efficiency of college teaching scoring.
当前的教育评价不仅受到简化、索引和数据化模式的限制，还受到高校教学评价科学性的限制。本工作首先对自然语言处理技术进行理论分析，分析智能评分的相关技术，设计高校英语教学智能评分的系统流程，最后对机器学习中的朴素贝叶斯算法进行理论研究。此外，还对高校英语教学智能评分的误差和评分分类的准确性进行分析和研究。结果表明，人工评分与机器评分之间的误差基本在 2 分左右，机器评分下高校英语教学的智能评分最小误差可达到 0 分。人工评分存在一定的偏差，机器评分可以减少这种误差的产生。朴素贝叶斯算法在高校智能评分数据集上具有最高的分类准确率，为 76.43%。加权朴素贝叶斯算法在高校英语教学智能评分的分类准确率上有所改进，平均准确率为 74.87%。综上所述，加权朴素贝叶斯算法在高校英语智能评分的分类准确率上具有更好的性能。本工作对自然语言处理下的高校智能教学评分系统的评分和朴素贝叶斯算法下的高校教学智能评分的分类具有显著效果，可提高高校教学评分的效率。
