Recover Missing Sensor Data with Iterative Imputing Network

TL;DR

This paper introduces an Iterative Imputing Network that effectively recovers missing sensor data by capturing complex temporal dynamics, significantly improving imputation accuracy over previous methods on benchmark datasets.

Contribution

The novel Iterative Imputing Network models latent temporal dynamics for sensor data imputation, outperforming existing interpolation-based methods on benchmark datasets.

Findings

Outperforms previous methods on Beijing air quality dataset

Maintains superiority across different missing data rates

Effectively captures complex temporal dynamics

Abstract

Sensor data has been playing an important role in machine learning tasks, complementary to the human-annotated data that is usually rather costly. However, due to systematic or accidental mis-operations, sensor data comes very often with a variety of missing values, resulting in considerable difficulties in the follow-up analysis and visualization. Previous work imputes the missing values by interpolating in the observational feature space, without consulting any latent (hidden) dynamics. In contrast, our model captures the latent complex temporal dynamics by summarizing each observation's context with a novel Iterative Imputing Network, thus significantly outperforms previous work on the benchmark Beijing air quality and meteorological dataset. Our model also yields consistent superiority over other methods in cases of different missing rates.