Rehman Ubaid Ur, Khan Meraj Ali, Al-Dayel Ibrahim, Mahmood Tahir
Department of Mathematics, University of Management and Technology, C-II, Johar Town, Lahore, 54700, Punjab, Pakistan.
Department of Mathematics and Statistics, College of Science, Imam Mohammad Ibn Saud Islamic University (IMSIU), 11566, Riyadh, Saudi Arabia.
Sci Rep. 2025 Jun 1;15(1):19195. doi: 10.1038/s41598-025-01909-z.
The selection of suitable AI models to predict disability diseases stands as a vital multi-attribute decision-making (MADM) task within healthcare technology. The current selection methods fail to integrate the management of uncertainties with bipolarity while also handling additional fuzzy information and linguistic terms during decision-making which leads to inferior model choices. To address these limitations, this paper proposes a new MADM approach within the environment of bipolar complex fuzzy linguistic sets (BCFLSs). In this manuscript, our primary contributions include, the proposal of four new Maclaurin symmetric mean (MSM) operators, in the setting of BCFLSs, analysis of properties of these operators to build the theoretical framework, development of a novel MADM approach to address uncertainties, bipolarity (dual aspects), addition fuzzy information; and linguistic terms (LTs), and application of the interpreted methodology to handle a real-life case study containing AI model selection for predicting disability disease. The case study of disability disease prediction results shows TensorFlow Neural Network achieved superior performance than other AI models with a score value of 7.776 using bipolar complex fuzzy linguistic MSM (BCFLMSM) and 1.943 using bipolar complex fuzzy linguistic weighted MSM (BCFLWMSM) operators while Support Vector Machine delivered the highest score (0.44 with bipolar complex fuzzy linguistic dual MSM (BCFLDMSM) and 0.006 with bipolar complex fuzzy linguistic weighted dual MSM (BCFLWDMSM) operators) based on different attribute interrelationships. Comparing the presented approach with the existing methodologies shows that the proposed approach is more efficient for handling complex decision situations. The findings suggest that our method offers more robust and accurate assessments by taking into account different aspects of uncertainty and system intricacy in the decision-making context.
选择合适的人工智能模型来预测残疾疾病是医疗技术领域一项至关重要的多属性决策（MADM）任务。当前的选择方法未能在决策过程中整合不确定性管理与双极性，同时处理额外的模糊信息和语言术语，这导致模型选择不佳。为了解决这些局限性，本文提出了一种在双极复杂模糊语言集（BCFLSs）环境下的新MADM方法。在本论文中，我们的主要贡献包括：在BCFLSs环境下提出四个新的麦克劳林对称均值（MSM）算子；分析这些算子的性质以构建理论框架；开发一种新颖的MADM方法来处理不确定性、双极性（两个方面）、附加模糊信息和语言术语；以及应用所阐述的方法来处理一个包含预测残疾疾病的人工智能模型选择的实际案例研究。残疾疾病预测结果的案例研究表明，使用双极复杂模糊语言MSM（BCFLMSM）时，TensorFlow神经网络的性能优于其他人工智能模型，得分为7.776，使用双极复杂模糊语言加权MSM（BCFLWMSM）算子时得分为1.943，而支持向量机在基于不同属性相互关系的情况下得分最高（使用双极复杂模糊语言对偶MSM（BCFLDMSM）时为0.44，使用双极复杂模糊语言加权对偶MSM（BCFLWDMSM）算子时为0.006）。将所提出的方法与现有方法进行比较表明，该方法在处理复杂决策情况时更有效。研究结果表明，我们的方法通过考虑决策背景中不确定性和系统复杂性的不同方面，提供了更稳健和准确的评估。
