Imputing missing values in sensor networks using sparse data representations

Wong, Liang Ze; Chen, Huiling; Lin, Shaowei; Chen, Daniel Chongli

doi:10.1145/2641798.2641816

Cited by 17 publications

(5 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is straightforward to implement and simple to understand MLA. KNN is one of the top MLAs [80], and data imputation algorithms based on KNN are widely used [81][82][83][84]. More complex and computationally extensive supervised learning algorithms than KNN, such as SVM, are also widely used for data imputation and handle both linear and nonlinear data efficiently [80,85].…”

Section: Data Imputationmentioning

confidence: 99%

Machine Learning-Enabled Internet of Things (IoT): Data, Applications, and Industry Perspective

Bzai

Alam²,

Dhafer

et al. 2022

Electronics

View full text Add to dashboard Cite

Machine learning (ML) allows the Internet of Things (IoT) to gain hidden insights from the treasure trove of sensed data and be truly ubiquitous without explicitly looking for knowledge and data patterns. Without ML, IoT cannot withstand the future requirements of businesses, governments, and individual users. The primary goal of IoT is to perceive what is happening in our surroundings and allow automation of decision-making through intelligent methods, which will mimic the decisions made by humans. In this paper, we classify and discuss the literature on ML-enabled IoT from three perspectives: data, application, and industry. We elaborate with dozens of cutting-edge methods and applications through a review of around 300 published sources on how ML and IoT work together to play a crucial role in making our environments smarter. We also discuss emerging IoT trends, including the Internet of Behavior (IoB), pandemic management, connected autonomous vehicles, edge and fog computing, and lightweight deep learning. Further, we classify challenges to IoT in four classes: technological, individual, business, and society. This paper will help exploit IoT opportunities and challenges to make our societies more prosperous and sustainable.

show abstract

Section: Data Imputationmentioning

confidence: 99%

Machine Learning-Enabled Internet of Things (IoT): Data, Applications, and Industry Perspective

Bzai

Alam²,

Dhafer

et al. 2022

Electronics

View full text Add to dashboard Cite

show abstract

“…1. https://github.com/pytorch/fairseq/tree/main/examples/wav2vec 2. https://huggingface.co/microsoft/deberta-large 3. https://github.com/zengqunzhao/MA-Net AE [66] is widely utilized in incomplete multimodal learning [33], [67]. It leverages autoencoders to impute missing data from partially observed input.…”

Section: Baselinesmentioning

confidence: 99%

GCNet: Graph Completion Network for Incomplete Multimodal Learning in Conversation

Zheng¹,

Chen²,

Sun³

et al. 2022

Preprint

View full text Add to dashboard Cite

Conversations have become a critical data format on social media platforms. Understanding conversation from emotion, content, and other aspects also attracts increasing attention from researchers due to its widespread application in human-computer interaction. In real-world environments, we often encounter the problem of incomplete modalities, which has become a core issue of conversation understanding. To address this problem, researchers propose various methods. However, existing approaches are mainly designed for individual utterances or medical images rather than conversational data, which cannot exploit temporal and speaker information in conversations. To this end, we propose a novel framework for incomplete multimodal learning in conversations, called "Graph Complete Network (GCNet)", filling the gap of existing works. Our GCNet contains two well-designed graph neural network-based modules, "Speaker GNN" and "Temporal GNN", to capture temporal and speaker information in conversations. To make full use of complete and incomplete data in feature learning, we jointly optimize classification and reconstruction in an end-to-end manner. To verify the effectiveness of our method, we conduct experiments on three benchmark conversational datasets. Experimental results demonstrate that our GCNet is superior to existing state-of-the-art approaches in incomplete multimodal learning.

show abstract

“…A more advanced approach is based on the autoencoder as a specific kind of neural network aiming to reconstruct inputs on its outputs. Here, one of the most commonly used models is the denoising autoencoder (DAE) [6], e.g., [16,17,18,19,20]. Typically, they are used in a discriminative way (see [19] for difference), meaning they impute a single value, which is deterministic once the network is trained.…”

Section: Related Workmentioning

confidence: 99%

Missing Features Reconstruction Using a Wasserstein Generative Adversarial Imputation Network

Friedjungová,

Vašata,

Balatsko

et al. 2020

Preprint

View full text Add to dashboard Cite

Missing data is one of the most common preprocessing problems. In this paper, we experimentally research the use of generative and non-generative models for feature reconstruction. Variational Autoencoder with Arbitrary Conditioning (VAEAC) and Generative Adversarial Imputation Network (GAIN) were researched as representatives of generative models, while the denoising autoencoder (DAE) represented non-generative models. Performance of the models is compared to traditional methods k-nearest neighbors (k-NN) and Multiple Imputation by Chained Equations (MICE). Moreover, we introduce WGAIN as the Wasserstein modification of GAIN, which turns out to be the best imputation model when the degree of missingness is less than or equal to 30%. Experiments were performed on real-world and artificial datasets with continuous features where different percentages of features, varying from 10% to 50%, were missing. Evaluation of algorithms was done by measuring the accuracy of the classification model previously trained on the uncorrupted dataset. The results show that GAIN and especially WGAIN are the best imputers regardless of the conditions. In general, they outperform or are comparative to MICE, k-NN, DAE, and VAEAC.

show abstract

Imputing missing values in sensor networks using sparse data representations

Cited by 17 publications

References 5 publications

Machine Learning-Enabled Internet of Things (IoT): Data, Applications, and Industry Perspective

Machine Learning-Enabled Internet of Things (IoT): Data, Applications, and Industry Perspective

GCNet: Graph Completion Network for Incomplete Multimodal Learning in Conversation

Missing Features Reconstruction Using a Wasserstein Generative Adversarial Imputation Network

Contact Info

Product

Resources

About