Application of EMD-LSTM Model in the Prediction of Tuberculosis Incidence in Shanxi Province
10.11783/j.issn.1002-3674.2025.03.003
- VernacularTitle:EMD-LSTM模型在山西省肺结核发病率预测中的应用
- Author:
Ruiqing ZHAO
1
;
Jing LIU
1
;
Zhiyang ZHAO
1
Author Information
1. 山西医科大学公共卫生学院卫生统计学教研室(030001)
- Publication Type:Journal Article
- Keywords:
Tuberculosis;
Long short-term memory;
Empirical mode decomposition;
Singular spectrum analysis;
Forecasting
- From:
Chinese Journal of Health Statistics
2025;42(3):334-339
- CountryChina
- Language:Chinese
-
Abstract:
Objective This study aimed to explore the feasibility of using a long short-term memory(LSTM)model based on empirical mode decomposition(EMD)and singular spectrum analysis(SSA)to predict the incidence of pulmonary tuberculosis in Shanxi Province.The goal was to provide a reliable prediction method to support the prevention and control of tuberculosis epidemics in the region.Methods Collecting and collating monthly data on the reported incidence of tuberculosis nationwide from January 2007 to December 2018.The LSTM、EMD-LSTM、SSA-LSTM models were established using the reported monthly incidence of tuberculosis reported in Shanxi Province from January 2007 to December 2017 as the training set and using the reported monthly incidence of tuberculosis from January to December 2018 as the test set.Mean squared error(MSE),mean absolute error(MAE),root mean squared error(RMSE),and mean absolute percentage error(MAPE)were used to evaluate the prediction effect of the models to determine the best model.Results The MSE,MAE,RMSE and MAPE of the EMD-LSTM model in predicting the incidence trend of pulmonary tuberculosis in the next year were 0.036,0.140,0.189 and 0.045,respectively.Compared with the LSTM model,the prediction performance increased by 66.36%,38.33%,42.38%and 41.56%,respectively.Compared with the SSA-LSTM model,it improved by 28.00%,9.68%,15.25%and 16.67%,respectively.Conclusion Compared with the single LSTM model,the fitting and prediction performance of EMD-LSTM and SSA-LSTM models are improved effectively.However,the prediction effect of EMD-LSTM model is better than that of SSA-LSTM model.Therefore,the EMD-LSTM model is more suitable for predicting the incidence trend of pulmonary tuberculosis in Shanxi Province,and can provide a theoretical basis for tuberculosis prevention and control policies.
