Integrated Energy System Evaluation Based on Missing Value Prediction and Index Correlation Analysis

The increasing demand for energy and the escalating environmental challenges necessitate urgent advancements in energy efficiency and the transformation of energy systems. The integrated energy system (IES), with its unique advantages, offers new pathways for sustainable energy development. In this context, this study proposes an evaluation method for IES that integrates back propagation neural networks (BPNNs) and self-organizing maps (SOMs). First, the BPNN is employed to predict missing values in the IES evaluation matrix, thereby reconstructing a complete dataset. The SOM is then used to cluster indexes, addressing interindex correlations, while the standard deviation method assigns weights to cluster centers. The individual index weights are derived from their deviations from the cluster centers, and IES schemes are ranked accordingly. Experiments based on the dataset of the IES of Xinxiang Pingyuan Hospital in Henan Province demonstrate that the proposed method is effective and outperforms several commonly used IES evaluation methods. This study presents an innovative approach for addressing missing data in evaluation matrices, thereby enhancing the robustness of IES evaluations under conditions of incomplete information. Additionally, the proposed clustering-based weighting mechanism improves the scientific rigor of index weighting, facilitating more systematic and reliable decision-making in complex IES evaluations. Furthermore, the research offers a practical reference for local governments in IES planning and contributes to advancing the theoretical framework of IES evaluation under incomplete data and interrelated indexes.