Year: 2024
Type: conference-paper
Abstract: E-commerce expansion and subsequent restrictions have significantly affect-ed consumer behaviour and the global economy, especially in the situation of COVID-19. The effects of those restrictions on the economy have been wide ranging, disrupting the supply chain and causing dips in consumer activity and falling corporate profits. In order to analyze the economic problems caused by restrictions on online shopping during this period, this study presents a novel deep learning-based strategy. The proposed approach com-bines a modified LSTM unit with the common time series prediction pipe-line, introducing an adaptive attention mechanism into the state-of-art LSTM architecture. This will improve the accuracy for identifying important trends and smaller changes in economic data. The model includes a 50* input which includes consumer spending, sales data, and macroeconomic indica-tors, allowing for a holistic analysis and forecasting. The adaptive attention mechanism allows for the model to dynamically determine and weight the most relevant points of data to utilize, enhancing both predictive power and interpretability of the model. Multiple case studies is the basis for the experimental analysis, with the aim of validating the model's ability to simulate and measure the impact of restrictions on online shopping on economic de-clines.
Author: Xiyue Zhu
Language: English
Topic: COVID-19 Pandemic Impacts
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W4410250708