Research on Evaluating the Digital Application Efficiency of Dongguan SMEs Using Random Forest Algorithm

In the context of accelerating global digitization, SMEs are increasingly faced with the need for digital transformation to improve operational efficiency and market competitiveness. However, current digital application efficiency assessment methods for SMEs generally suffer from incomplete dimensionality and lack of objectivity, and are especially limited in dealing with multidimensional data and few analysis categories. To address these challenges, this study constructs a digital efficiency assessment model based on the random forest algorithm to systematically analyze the performance of Dongguan SMEs in the process of digital transformation. The model screens out key variables through a recursive feature elimination method, and combines data standardization and parameter optimization strategies to achieve efficient classification. In the experiment, the model performs well in the core indicators of classification accuracy (89%), precision (85%), and recall (83%), and effectively discriminates high-, medium-, and low-efficiency enterprises. In the analysis results of stacked bar charts, different categories of enterprises show significant differences in digital application efficiency, further revealing the critical impact of digital tools and technological capabilities on enterprise performance. This study not only provides a scientific and quantitative analysis tool to assess the digital efficiency of SMEs, which helps enterprises optimize their digital transformation path, but also provides data support and theoretical basis for local governments to formulate precise support policies.