Robustness Analysis of Intelligent Development: Replacing Explanatory Variables and Alternative Measurement Approach

To enhance the robustness of our analysis, we replaced the explanatory variables. Specifically, we utilized the entropy weight method in benchmark regression and employed vertical and horizontal methods in this aspect to establish the comprehensive evaluation system of intelligent development. The regression results are presented in columns (1) and (2) of Table 4. Furthermore, we retrieved information on the development of enterprise-level intelligence from the 'Enterprise Query' platform, utilizing Python crawler technology to search for contents related to intelligent technologies such as 'intelligence', 'cloud', 'data', 'Internet', and 'machine learning' in the columns of 'business scope' and 'company name'. The number of intelligent enterprises in each city was then utilized to measure the level of intelligent development, and the results are reported in columns (3) and (4).