%0 Journal Article
%T 基于深度森林的在线课程购买行为预测研究<br>Research on Online Course Purchase Behavior Prediction Based on Deep Forest
%A 胡陈陈
%A 吕卫东
%A 郑江怀
%A 王一朵
%J Advances in Applied Mathematics
%P 4306-4312
%@ 2324-8009
%D 2022
%I Hans Publishing
%R 10.12677/AAM.2022.117457
%X 当前疫情环境下的生活条件与人们日益增加的教育需求，不断地推动着各种类型在线教育的发展。在传统机器学习对在线课程用户行为预测的基础上，本文采用深度学习中深度森林方法对用户的购买行为进行预测，选取准确率(Accuracy)、精度(Precision)、召回率(Recall)以及F1值作为模型评价指标，比较不同模型的预测精度。针对预测准确率最高的深度森林模型，进行多次参数调优得到最优参数模型，模型的最终准确率为98.744%。采用深度森林模型对用户购买预测，为企业进行精准营销、减少宣传投入提供重要的参考依据。<br />
The living conditions in the current epidemic environment and the increasing educational needs of people are constantly promoting the development of various types of online education. On the basis of traditional machine learning to predict online course users’ behavior, this paper uses the deep forest method in deep learning to predict users’ purchasing behavior, and selects accuracy, preci-sion, recall and F1 value as model evaluation indicators to compare the prediction accuracy of dif-ferent models. Aiming at the deep forest model with the highest prediction accuracy, the optimal parameter model is obtained by multiple parameter tuning. The final accuracy of the model is 98.744%. The deep forest model is used to predict the purchase of users, which provides an im-portant reference basis for enterprises to carry out precision marketing and reduce publicity in-vestment.
%K 在线课程，购买行为预测，随机森林，深度森林<br>Online Courses
%K Prediction of Purchase Behavior
%K Random Forest
%K Deep Forest
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=53364