基于数据包络分析与多目标软子空间聚类的房地产企业运营风险综合评价研究

doi:10.12005/orms.2025.0182

摘要/Abstract

摘要： 本研究基于数据包络分析和多目标软子空间聚类方法,对房地产企业运营风险开展综合评价研究,为企业决策者提供风险管理建议,以降低企业可能面临的损失。首先,本文采用多目标软子空间聚类法对房地产企业风险进行聚类。在此基础上,通过与多个传统聚类算法的对比,验证了多目标软子空间聚类算法的适用性。结果表明,该方法相较于传统聚类方法结果更准确,在高维数据上表现较好,并且能够获得不同维度下各风险维度的权重值。其次,运用超效率数据包络分析模型对房地产企业的运营风险评估。最后,构建运营风险评价体系,分析了房地产企业的风险来源。研究结果表明,司法风险、关联风险类型对房地产企业运营具有较大影响;不同风险等级的企业存在不同的风险来源,应采取不同的风险管理措施。因此,本研究根据上述结果,为房地产企业决策者提供有效的指导建议,以期降低企业经营风险,增加企业的稳定性和抵御风险的能力,提高房地产企业运营风险管理水平,为企业的长期稳定发展提供可靠保障。

关键词: 数据包络分析, 软子空间聚类, 房地产, 运营风险评估

Abstract: In recent years, the Chinese real estate market has been faced with challenges, such as rapidly surging housing prices and the accumulation of corporate debt risks, leading to substantial operational risks for real estate firms. Consequently, the accurate assessment of operational risks associated with real estate firms has significant theoretical implications and practical value. This study seeks to provide a comprehensive evaluation of the operational risks of real estate firms by basing its methodology on Data Envelopment Analysis (DEA) and a multi-objective soft subspace clustering approach. The aim is to equip corporate decision-makers with valuable risk management insights, which can help in mitigating potential losses.
The first stage of the research involves using a multi-objective soft subspace clustering technique to categorize the risks associated with real estate firms. This method is chosen after a comparative analysis with multiple traditional clustering algorithms, which validates its applicability. The results indicate that this approach provides an accurate risk assessment compared to traditional clustering methods. It also performs well on high-dimensional data, can effectively extract differential features of data clusters even in sparse data situations, and can obtain weight values for different risk dimensions under various dimensions. Following the clustering phase, the study employs a super-efficiency DEA model to further evaluate the operational risks of real estate firms. The model’s efficiency in risk evaluation offers a solid foundation for a robust risk assessment framework. As an extension of the traditional CCR model, the super-efficiency model provides more discriminating power in differentiating the efficient units, thus permitting a more nuanced understanding of the operational risks inherent in real estate firms. Building on these assessments, the research constructs an operational risk evaluation system and delves into an analysis of the sources of risk for real estate firms. The findings reveal that legal risks and associated risk types have substantial impacts on the operations of real estate firms. Distinct risk sources exist for businesses with different risk ratings, indicating the necessity of adopting tailored risk management strategies. The results of this study offer effective guidance for decision-makers in real estate firms, with the goal of reducing business operating risks, enhancing the stability and risk resistance of enterprises, and improving the level of operational risk management in real estate firms. By providing these insights, the research aims to contribute to a more reliable safeguard for the long-term steady development of enterprises.
In conclusion, this study presents a novel approach to risk assessment in the real estate sector, combining the use of multi-objective soft subspace clustering and super-efficiency DEA. This dual-method approach not only enables more granular analysis of risk factors but also provides actionable insights for firms to manage their operational risks effectively. The proposed operational risk evaluation system serves as a valuable tool for firms to understand their risk profiles better and formulate appropriate strategies to mitigate these risks. This research, therefore, makes both theoretical and practical contributions to the field of operational risk management in the real estate industry.
Further research in this area could branch into several directions: (1)Utilize other types of DEA models like the Undesirable Outputs super-slacks-based measure model and the benevolent cross-efficiency model. By comparing the results, we could explore early warning mechanisms for corporate operational risks. (2)Apply this model to the evaluation of operational risks in industries other than real estate. This would further refine the model design and the construction of the corporate operational risk assessment system. (3)Further improve the heuristic algorithms and objectives used in the subspace clustering model. The goal is to converge in a shorter time and obtain relatively effective and fair results. These measures will ensure the continuous evolution and adaptation of the comprehensive risk assessment model to the ever-changing business environment, thereby providing a solid safeguard for risk assessment and risk mitigation.

Key words: data envelopment analysis, soft subspace clustering, real estate companies, operational risk assessment

中图分类号:

F832.4

赵婧, 戴浥, 高显. 基于数据包络分析与多目标软子空间聚类的房地产企业运营风险综合评价研究[J]. 运筹与管理, 2025, 34(6): 107-114.

ZHAO Jing, DAI Yi, GAO Xian. Comprehensive Evaluation of Enterprise Operation Risk Based on Data Envelopment Analysis and Multi-objective Soft Subspace Clustering[J]. Operations Research and Management Science, 2025, 34(6): 107-114.

参考文献

[1] 郭娜,王少严,胡佳琪.房地产价格、金融稳定与宏观审慎监管—基于NK-DSGE模型的研究[J].武汉金融,2022(6):3-12.
[2] 杨源源,贾鹏飞,高洁超.中国房地产长效调控范式选择:房产税政策还是宏观审慎政策[J].财贸经济,2021,42(8):53-66.
[3] 冯明.国民经济核算视角下中国居民消费率的因素分解研究—对“消费能力说”和“消费意愿说”的定量考察[J].数量经济技术经济研究,2023,40(5):180-201.
[4] ELLINGTON M, FU X, ZHU Y. Real estate illiquidity and returns: A time-varying regional perspective[J]. International Journal of Forecasting, 2023, 39(1): 58-72.
[5] 代婉瑞,宋良荣.商业银行数字化转型与信用风险经济资本管理:研究综述与展望[J].财会月刊,2022(15):130-137.
[6] XIAO P, SALLEH M I, ZAIDAN B B, et al. Research on risk assessment of blockchain-driven supply chain finance: A systematic review[J]. Computers & Industrial Engineering, 2023,176: 108990.
[7] WU D D, OLSON D. Enterprise risk management: A DEA VaR approach in vendor selection[J]. International Journal of Production Research, 2010, 48(16): 4919-4932.
[8] CHARNES A, COOPER W W, RHODES E. Measuring the efficiency of decision making units[J]. European Journal of Operational Research, 1978, 2(6): 429-444.
[9] LIU J S, LU L Y Y, LU W M. Research fronts in data envelopment analysis[J]. Omega, 2016, 58: 33-45.
[10] ANDERSEN P, PETERSEN N C. A procedure for ranking efficient units in data envelopment analysis[J]. Management Science, 1993, 39(10): 1261-1264.
[11] 罗春婵,尤秋爽,何代弟.公司治理异质性与商业银行风险约束效率[J].金融与经济,2021(3):21-29.
[12] SHI S, TSE R, LUO W, et al. Machine learning-driven credit risk: A systemic review[J]. Neural Computing & Applications, 2022, 34(17): 14327-14339.
[13] DUAN C Y, CHEN X Q, SHI H, et al. A new model for failure mode and effects analysis based on k-means clustering within hesitant linguistic environment[J]. IEEE Transactions on Engineering Management, 2019, 69(5): 1837-1847.
[14] OSMAN A, SHEHADEH M. Risk assessment of interstate pipelines using a fuzzy-clustering approach[J]. Scientific Reports, 2022, 12(1): (Article)13750.
[15] D’URSO P. Informational Paradigm, management of uncertainty and theoretical formalisms in the clustering framework: A review[J]. Information Sciences, 2017, 400: 30-62.
[16] ZHU Y, TING K M, CARMAN M J. Grouping points by shared subspaces for effective subspace clustering[J]. Pattern Recognition, 2018, 83: 230-244.
[17] LIU C, XIE J, ZHAO Q, et al. Novel evolutionary multi-objective soft subspace clustering algorithm for credit risk assessment[J]. Expert Systems with Applications, 2019, 138: 112827.
[18] MA S, LYU S, ZHANG Y. Weighted clustering-based risk assessment on urban rainstorm and flood disaster[J]. Urban Climate, 2021, 39: 100974.
[19] SONG Y, WANG Y, YE X, et al. Multi-view ensemble learning based on distance-to-model and adaptive clustering for imbalanced credit risk assessment in P2P lending[J]. Information Sciences, 2020, 525: 182-204.
[20] 周开乐,杨善林,丁帅,等.聚类有效性研究综述[J].系统工程理论与实践,2014,34(9):2417-2431.
[21] PARVIN H, MINAEI-BIDGOLI B. A clustering ensemble framework based on elite selection of weighted clusters[J]. Advances in Data Analysis and Classification, 2013, 7: 181-208.