A Fair Ranking Mechanism for Multi-Source Evaluation Based on Random Forest and RMF-Ranking
DOI:
https://doi.org/10.62051/rrttqq60Keywords:
RMF-Ranking Algorithm; Random Forest; SHAP Interpretability Analysis.Abstract
This paper investigates the fairness and stability of competition rankings in multi-source evaluation scenarios. We construct a comprehensive framework that combines audience voting reconstruction, counterfactual simulation, and analysis using random forests and SHAP explanations, and propose the RMF-Ranking fairness mechanism. This method restores unobservable voting distributions through macro- and micro-level constraints, compares the impact of different scoring mechanisms on ranking results, and further identifies structural differences between judge evaluations and audience preferences. Building on this foundation, RMF-Ranking transforms diverse evaluation results into a ranking aggregation problem and employs a Markov approximation to derive a consensus ranking, effectively mitigating the excessive dominance of any single evaluation dimension. Research indicates that this mechanism reduces extreme ranking fluctuations, enhances result robustness, and is suitable for fair evaluation and ranking decision-making scenarios involving multiple stakeholders and multiple metrics.
Downloads
References
[1] Guo Kewei, Ren Guojun, Jia Jiyuan. Assessment of the Resilience of the Arable Land System in Yunxi County and Analysis of Its Driving Factors Based on Random Forests and SHAP [J]. Central South Agricultural Science and Technology, 2025, 46(S1): 145–150.
[2] Yin Zhichao, Liu Jiayi, Li Feng. The Impact of the New Individual Income Tax Policy on Household Consumption: A Counterfactual Analysis Based on a Nonlinear Consumption Function [J]. Economic Research Journal, 2026, 61(2): 213-236.
[3] Ren Haoyuan, Jin Jing, Wang Yanyan, et al. Data Mining of Temperature and Humidity Data in Coastal Areas Based on the K-Means Clustering Algorithm and Bayesian Network Model [J]. Aerospace Environmental Engineering, 2025, 42(5): 494-503.
[4] Tan Zhicheng. A Study on the Determinants of Commercial Housing Prices in Qingyuan City Based on the Random Forest Model [D]. Guangdong University of Technology, 2025. DOI: 10.27029/d.cnki.ggdgu.2025.002458.
[5] Dong Hao, Wang Hefei, Lei Jiaqi. A Markov Chain-Based Passenger Flow Allocation Model for Tourism Rail Transit [J]. Research on Urban Rail Transit, 2022, 25(9): 38-44. DOI:10.16037/j.1007-869x.2022.09.008.
[6] Xiao Fuxian. Construction and Stability Analysis of the Chinese Stock Market Correlation Network from an Evolutionary Network Perspective [D]. Wuhan Textile University, 2023. DOI:10.27698/d.cnki.gwhxj.2023.000758.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Transactions on Computer Science and Intelligent Systems Research
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
