MBDiff: Multi-view Behavior-aware Diffusion Model for Probabilistic Utility Data Imputation

2026-07-31

AI Summary

This study addresses the challenge of extensive missing data in utility records—such as electricity, water, and gas consumption—caused by equipment failures and transmission issues, which compromise billing accuracy and supply–demand management efficiency. To tackle this problem, the authors propose MBDiff, a novel model that uniquely integrates multi-view user behavior modeling with a conditional diffusion mechanism. By capturing global, local, and instance-level behavioral patterns and incorporating a behavior-aware reference selection strategy alongside a conditional attention-based denoising network, MBDiff enables effective probabilistic imputation of missing values. Experimental results on a large-scale municipal dataset from Florida demonstrate that MBDiff substantially outperforms existing methods, achieving improvements of 7.04% and 29.1% in imputation accuracy for missing blocks of electricity and water usage data, respectively.

Abstract

Utility data (e.g., electricity, water, and gas consumption), collected by ubiquitous sensors and embedded devices, often contains substantial missing values due to various factors such as device failures and data transmission issues. The data missingness can severely impact utility billing accuracy, hinder demand forecasting, and disrupt efficient utility supply management. As a result, utility data imputation has attracted much interest from both industry and academia. While many studies have attempted to address this issue, most of them rely on aggregated datasets for training, overlooking rich user behavior information, which could provide valuable insights for more accurate imputation. However, learning comprehensive user behavior from long-term, diverse, and incomplete utility data remains a significant challenge. Moreover, leveraging user behavior information to guide imputation is nontrivial due to the indirect nature of the correlations. To address these challenges, we propose MBDiff, a Multi-view Behavior-aware Diffusion Model for Probabilistic Utility Data Imputation. MBDiff incorporates two key technical components: (i) a multi-view User Behavior Extraction module that learns comprehensive user behavior from multiple perspectives, including global, local, and instance-level views; and (ii) a behavior-aware conditional diffusion model consisting of a reference selection module and a conditional attentional denoising network to impute utility data in a computationally efficient manner. We implement and evaluate MBDiff by collaborating with one of the largest municipal utility providers in Florida. Experimental results demonstrate our proposed MBDiff effectively outperforms state-of-the-art baselines, e.g., it improves 7.04% and 29.1% on the electricity and water usage datasets for block missingness imputation, respectively.

Problem

• utility data imputation
• missing data
• user behavior
• probabilistic imputation
• multi-view learning

Innovation

• multi-view behavior extraction
• conditional diffusion model
• utility data imputation
• behavior-aware attention
• probabilistic imputation