Abstract Funded pension systems are essential for ensuring retirement security but are increasingly exposed to market volatility, demographic changes, and economic uncertainty. Traditional risk assessment models often lack adaptability and interpretability, limiting their effectiveness in dynamic financial environments. To address this gap, this research proposes a hybrid Artificial Intelligence and fuzzy logic-based model, namely Sugeno Fuzzy Inference with Adaptive Long Short-Term Memory Network (SFI-ALSTMN), for pension finance risk management. The model is developed using a dataset comprising 4000 records with 17 financial, demographic, and macroeconomic features. Data preprocessing includes Min–Max normalization, K-Nearest Neighbors (KNN) imputation, and Interquartile Range (IQR)-based outlier handling, followed by feature reduction using Principal Component Analysis (PCA). The proposed framework is implemented using the Python programming environment for model development and evaluation. The ALSTMN component captures temporal financial patterns, while the SFI system enhances interpretability by managing uncertainty through rule-based reasoning. Experimental results demonstrate that the proposed model achieves improved predictive performance with Root Mean Squared Error (RMSE) of 0.7404, Mean Absolute Error (MAE) of 0.5214, Mean Absolute Scaled Error (MASE) of 0.5623, and Mean Squared Error (MSE) of 0.5482. Furthermore, the model attains a Coefficient of Determination (R²) value of 0.3527, indicating moderate but consistent prediction capability. The model supports more reliable and interpretable pension risk assessment, contributing to improved financial decision-making, risk mitigation, and long-term pension system sustainability.
Ya Xing (Sat,) studied this question.