Analysis and Prediction of Economic Cross-correlation under COVID-19 based on MF-LSTM and WNN

Abstract

In the era of COVID-19, it is particularly important to analyze the correlation of economic indicators and propose corresponding policies. In this paper, a number of industry indicators that have an important impact on the economy are selected, and normalization, interpolation, and PCA operations are performed on them. Based on the MF-LSTM neural network, this paper analyzes the many-to-one correlation between industry indicators and macroeconomic indicators. Furthermore, based on the WNN neural network, wavelet analysis is used to predict the impact of macroeconomic indicators on people's livelihood indicators under time series. Based on the above model, the coupling relationship between industry indicators and macroeconomic indicators and the development trend of people's livelihood indicators for a period of time in the future have been obtained, and the accuracy of the model has also been verified.