Construction of e-commerce major group course evaluation system based on CNN and LDA

Analyzing course evaluation data is a key means to improve the quality of course construction for e-commerce majors. To this end, a course evaluation system for e-commerce majors based on convolutional neural network (CNN) and latent Dirichlet allocation (LDA) is constructed to realize the emotional tendency analysis of course evaluation and the acquisition of evaluation subject words. In terms of data collection, python crawler technology is used to obtain data. In terms of data preprocessing, Python's deduplication method and regular expression operations are used to complete data cleaning, and jieba is used to implement Chinese evaluation word segmentation. In terms of emotional tendency analysis, Word2Vec is used to convert text data into Word Embeddings, train and test the CNN model, and finally apply it to the emotional tendency analysis task. In term of sentiment topic analysis, TF-IDF is used to calculate the keywords of the evaluation data, and the LDA model is constructed to obtain the topics and subject words of the evaluation. Through experiments, it is found that the course evaluation system constructed in the article can realize the course evaluation of any e-commerce major group courses.