Construction of a Model for Evaluating the Effects of Regional Ethnic Autonomy Laws and Regulations Based on Differentiated Big Data Analysis

This paper constructs a model for evaluating the effectiveness of laws and regulations on ethnic regional autonomy based on differentiated big data analysis. It innovatively integrates the Fuzzy Best-Worst Method (FBWM), the CRITIC objective weighting method, and the principle of minimum discriminative information to form the FBWM-CRITIC combined weighting model. First, the current state of legislation for autonomous regulations and single-issue regulations is systematically reviewed, establishing four primary indicators: legal effectiveness (A1), political effectiveness (A2), economic effectiveness (A3), and social effectiveness (A4). These are further broken down into 18 secondary indicators and over 50 tertiary observational indicators. Through combined weighting, the core weights were determined: social effects (A4) accounted for 35.1%, political effects (A2) for 23.9%, highlighting the critical importance of human rights protection and public services; economic effects (A3) accounted for only 20.2%, with the lowest weight (0.034) among sub-indicators being technological development (B12). Model validation indicates that the FBWM consistency ratio (CR = 0.041–0.059) significantly outperforms AHP (CR = 0.075–0.132), and the combined weighting error rate is 23.7% lower than traditional additive/multiplicative synthesis methods. Applying this model to evaluate the implementation effectiveness of the five autonomous regions from 2015 to 2024 reveals that all dimensions are at the “qualified” level (60–70 points), with economic effectiveness being the best (66.79 points) and legal effectiveness being the weakest (63.06 points). The overall effectiveness improved by 47.2% over the decade (from an average of 0.443 in 2015 to 0.652 in 2024), Xinjiang (0.708) and Ningxia (0.692) maintained their lead, while Tibet showed significant fluctuations (from 0.385 in 2015 to 0.584 in 2024). Markov chain analysis revealed that the probability of low-level regions being locked in reached 88%, while high-level regions (Category IV) achieved 100% stability and exerted positive spillover effects on neighboring regions (increasing the transfer probability by 12%).