Application of data mining technology in spatio-temporal analysis of the correlation between R&D investment and economic benefits in high-tech industries

In the context of the deep restructuring of the global economy, this paper focuses on the analysis of the temporal and spatial correlation between R&D investment and economic benefits in high-tech industries, aiming to reveal the optimization path of R&D resource allocation. In terms of research methods, this paper innovatively integrates spatial Durbin model (embedded regional interaction effect), dynamic lag weight calculation and spatio-temporal double fixed effect to construct a four-layer computing architecture. In addition, the model is verified based on the 32-country data sets of OECD and World Bank. The results show that the prediction accuracy of the model is significantly improved (R2=0.912, which is 19.6% higher than that of traditional OLS), the contribution of spatial spillover effect reaches 19.4%, labor quality is the core driving factor (SHAP value 32.7%), the error rate of policy simulation is ≤6%, and the parameter volatility under data disturbance in robustness test is <7%. The above results show that the framework provides a high-precision and anti-interference spatio-temporal quantitative decision-making tool for R&D investment benefits in high-tech industries.