Construction and Application of Macroeconomic Forecasting Model Based on Time Series Clustering Analysis

The current macroeconomic forecasting models cannot be fully applicable to all economic phenomena, and manual analysis is also difficult to comprehensively consider multiple factors. To address this issue, this article collected and preprocessed macroeconomic data, used a vector autoregression (VAR) model to process time series data, and calculated the Euclidean distance to measure the similarity between economic variables. Based on time series clustering analysis, input the clustering results into the model to predict future macroeconomic variables. Research has found that the correlation between GDP growth and multiple economic variables exhibits significant differences over specific time periods. The average absolute deviation (MAD) between the predicted Gross Domestic Product (GDP) and the actual GDP by the model is less than 650.6 billion yuan. This article demonstrates that macroeconomic forecasting models based on time series clustering analysis can improve decision-making accuracy and contribute to promoting economic stability and development.