Abstract:Objective To analyze and predict the capacity of rabies vaccination in Gansu Province by using long short-term memory (LSTM) and gate recurrent unit (GRU) neural network. Methods The daily rabies vaccination capacity in Gansu Province from January 1, 2020 to December 31, 2021 was collected to establish the models based on LSTM and GRU neural network respectively.Data about the daily rabies vaccination capacity from January 1, 2022 to August 30, 2022 were used to test and compare the prediction effects of the two models. Results From January 1, 2020 to August 30, 2022, a total of 893,123 doses of vaccines against rabies had been cumulatively administered. The root mean squared error (RMSE) of the rabies vaccination capacity from January 1, 2022 to August 30, 2022 predicted by the models based on LSTM and GRU neural network was 86.06 and 102.63 respectively, and the mean absolute error (MAE) was 73.62 and 83.11 respectively. The results indicated that the prediction accuracy of LSTM neural network was superior to that of GRU neural network, and the prediction results were basically consistent with the actual vaccination capacity. Conclusion By accurately predicting the capacity of rabies vaccination in Gansu Province, accurate planning can be provided for production and distribution enterprises of rabies vaccines so as to improve the efficiency of vaccine circulation and reduce the risk of excessive and inadequate production and storage of vaccines.
梁雪枫, 刘舒瑜, 那金, 焦永卓, 邓琳, 张晓曙. LSTM和GRU神经网络对甘肃省狂犬疫苗接种量的预测效果比较[J]. 实用预防医学, 2023, 30(12): 1453-1455.
LIANG Xuefeng, LIU Shuyu, NA Jin, JIAO Yongzhuo, DENG Lin, ZHANG Xiaoshu. Comparison of the prediction effect of LSTM and GRU neural network on the capacity of rabies vaccination in Gansu Province. , 2023, 30(12): 1453-1455.