Chen Bo-Lun, Shen Yi-Yun, Zhu Guo-Chang, Yu Yong-Tao, Ji Min
Department of Computer Science, Huaiyin Institute of Technology, Huaiyin, 223003 Jiangsu China.
Institute of Informatics, University of Zurich, 8050 Zurich, Switzerland.
Neural Process Lett. 2022 Apr 25:1-22. doi: 10.1007/s11063-022-10836-3.
At present, the Corona Virus Disease 2019 (COVID-19) is ravaging the world, bringing great impact on people's life safety and health as well as the healthy development of economy and society, so the research on the prediction of the development trend of the epidemic is crucial. In this paper, we focus on the prevention and control of epidemic using the relevant technologies in the field of artificial intelligence and signal analysis. With the unknown principle of epidemic transmission, we first smooth out the complex and variable epidemic data through the empirical mode decomposition model to obtain the change trends of epidemic data at different time scales. On this basis, the change trends under different time scales are trained using an extreme learning machine to obtain the corresponding prediction values, and finally the epidemic prediction results are obtained by fitting through Adaptive Network-based Fuzzy Inference System. The experimental results show that the algorithm has good learning ability, especially in the prediction of time-series sequences can guarantee the accuracy rate while having low time complexity. Therefore, this paper not only plays a theoretical support for epidemic prevention and control, but also plays an important role in the construction of public emergency health system in the long run.
目前，新型冠状病毒肺炎（COVID-19）正在肆虐全球，给人们的生命安全和健康以及经济社会的健康发展带来巨大影响，因此对疫情发展趋势预测的研究至关重要。本文利用人工智能和信号分析领域的相关技术聚焦于疫情防控。鉴于疫情传播原理未知，我们首先通过经验模态分解模型对复杂多变的疫情数据进行平滑处理，以获得不同时间尺度下疫情数据的变化趋势。在此基础上，利用极限学习机对不同时间尺度下的变化趋势进行训练以获得相应的预测值，最后通过基于自适应网络的模糊推理系统进行拟合得到疫情预测结果。实验结果表明，该算法具有良好的学习能力，尤其在时间序列预测方面能够在保证准确率的同时具有较低的时间复杂度。因此，本文不仅为疫情防控提供了理论支持，从长远来看也对公共应急卫生体系建设具有重要作用。
