摘要
保障高校食堂早餐科学合理供应是一项重要的民生工程。基于校园一卡通系统的消费数据,对高校食堂早餐进行了统计分类,采用基于长短时记忆网络(LSTM)的改进模型对早餐供应展开了研究,并对早点、炒饭、面条、粥、豆浆5种常见早餐进行了分类预测。试验结果表明,改进的LSTM模型对5个类别预测的均方根误差(RMSE)平均值为2.19,平均绝对误差(MAE)平均值为3.42;与自回归移动平均模型(ARAM)、循环神经网络(RNN)和门控循环单元(GRU)3个经典的时间序列模型相比,改进的LSTM模型表现最出色,具有较高的预测准确性和可靠性。
It is an important livelihood project to ensuring the scientific and reasonable supply of breakfast in university canteens.Based on the consumption data of the campus all-in-one card system,the breakfast in the university canteen is statistically classified,and the breakfast supply is studied by using an improved model based on the Long Short Memory Network(LSTM).Five common breakfasts,such as breakfast,fried rice,noodles,congee,and soybean milk,are classified and predicted.The experimental results show that the average root mean square error(RMSE)of the improved LSTM model for predicting five categories is 2.19,and the average absolute error(MAE)is 3.42.Compared with three classic time series models,such as Autoregressive Moving Average(ARAM),Recurrent Neural Network(RNN),and Gated Recurrent Unit(GRU),the improved LSTM model performs the best,with high prediction accuracy and reliability,providing an effective prediction model for university canteens.
作者
袁以铎
YUAN Yiduo(Department of Logistics Management and Infrastructure,Chuzhou University,Chuzhou Anhui 239000,China)
出处
《长春工程学院学报（自然科学版）》
2024年第1期107-111,共5页
Journal of Changchun Institute of Technology：Natural Sciences Edition
关键词
高校食堂早餐
长短时记忆网络
校园一卡通
university canteen breakfast
long short-term memory
campus all-in-one card
