Causal View of Time Series Imputation: Some Identification Results on Missing Mechanism

2025-05-12

AI Summary

Time-series imputation faces challenges from heterogeneous missingness mechanisms; existing methods overlook the fundamental distinction between missing-at-random (MAR) and missing-not-at-random (MNAR), leading to mechanism mismatch and biased imputations. This paper proposes a mechanism-aware causal generative model. First, it establishes identifiability theory for latent variables under arbitrary missingness mechanisms within a nonlinear independent component analysis framework. Second, it designs a dual-path architecture integrating normalizing flows and variational inference to jointly learn latent temporal dynamics and missingness cause variables, augmented by an explicit causal graph encoding the missingness mechanism. Third, it introduces a mechanism-customized imputation paradigm. Evaluated across diverse real-world datasets, our method consistently outperforms state-of-the-art baselines, reducing imputation error by over 23% under MNAR conditions—demonstrating both the effectiveness and generalizability of mechanism-adapted modeling.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Relational Probabilistic Models
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Time series imputation is one of the most challenge problems and has broad applications in various fields like health care and the Internet of Things. Existing methods mainly aim to model the temporally latent dependencies and the generation process from the observed time series data. In real-world scenarios, different types of missing mechanisms, like MAR (Missing At Random), and MNAR (Missing Not At Random) can occur in time series data. However, existing methods often overlook the difference among the aforementioned missing mechanisms and use a single model for time series imputation, which can easily lead to misleading results due to mechanism mismatching. In this paper, we propose a framework for time series imputation problem by exploring Different Missing Mechanisms (DMM in short) and tailoring solutions accordingly. Specifically, we first analyze the data generation processes with temporal latent states and missing cause variables for different mechanisms. Sequentially, we model these generation processes via variational inference and estimate prior distributions of latent variables via normalizing flow-based neural architecture. Furthermore, we establish identifiability results under the nonlinear independent component analysis framework to show that latent variables are identifiable. Experimental results show that our method surpasses existing time series imputation techniques across various datasets with different missing mechanisms, demonstrating its effectiveness in real-world applications.

Problem

• Addressing missing mechanisms in time series imputation
• Identifying latent variables for different missing types
• Improving accuracy in MAR and MNAR scenarios

Innovation

• Explores Different Missing Mechanisms (DMM) for imputation
• Uses variational inference and normalizing flow-based architecture
• Establishes identifiability under nonlinear independent component analysis