摘要
医院门诊量本质上是一种具有潜在规律的时间序列,通过对门诊量进行有效分析和预测,可以更加科学、合理地配置医疗资源。针对门诊量波动幅度较大的时间序列预测问题,提出CEEMDAN-VMD-SSA-LSTM模型。通过完全自适应噪声完备集合经验模态分解(CEEMDAN)和变分模态分解(VMD)对数据进行两次经验模态分解,提高门诊量数据集的准确性和稳定性。采用在时序问题处理上具有良好性能的长短期记忆(LSTM)神经网络,并通过寻优能力强、稳定性好的麻雀搜索算法(SSA)对LSTM网络超参数进行优化,得到预测模型。通过比较实验,提出方法可以更加精准地对门诊量进行预测和分析,为医院更好地运营管理提供了重要依据和决策支持。
Hospital outpatient volume is essentially a time series with potential laws.Through effective analysis and prediction of outpatient volume,medical resources can be more scientifically and reasonably allocated.A CEEMDAN-VMD-SSA-LSTM model is proposed to predict the time series with large fluctuation of outpatient volume.The data are decomposed twice by fully adaptive noise complete set empirical mode decomposition(CEEMDAN)and variational mode decomposition(VMD),which improves the accuracy and stability of the outpatient data set.The long short-term memory(LSTM)neural network with good performance in time series problem processing is adopted,and the super parameters of LSTM network are optimized by sparrow search algorithm(SSA)with strong optimization ability and good stability.Through comparative experiments,the proposed method can more accurately predict and analyze the outpatient volume,providing an important basis and decision support for better operation and management of the hospital.
作者
樊冲
FAN Chong(Jinzhou Big Data Management Center,Jinzhou 121000,China)
出处
《微型电脑应用》
2024年第5期214-218,242,共6页
Microcomputer Applications
关键词
自适应噪声完备集合经验模态分解
变分模态分解
长短记忆网络
麻雀搜索算法
adaptive noise complete set empirical mode decomposition
variational modal decomposition
long short-term memory network
sparrow search algorithm
