Filters








1,372 Hits in 4.9 sec

Uncertainty-Aware Variational-Recurrent Imputation Network for Clinical Time Series [article]

Ahmad Wisnu Mulyadi, Eunji Jun, Heung-Il Suk
2020 arXiv   pre-print
In this work, we propose a novel variational-recurrent imputation network, which unifies an imputation and a prediction network by taking into account the correlated features, temporal dynamics, as well  ...  Specifically, we leverage the deep generative model in the imputation, which is based on the distribution among variables, and a recurrent imputation network to exploit the temporal relations, in conjunction  ...  We propose a novel variational-recurrent imputation network (V-RIN) which unifies the imputation and prediction networks for multivariate time series EHR data, governing both correlations among variables  ... 
arXiv:2003.00662v2 fatcat:gkm463rpangm5l7fx7itmaocye

Path Imputation Strategies for Signature Models of Irregular Time Series [article]

Michael Moor, Max Horn, Christian Bock, Karsten Borgwardt, Bastian Rieck
2020 arXiv   pre-print
Next, we observe that uncertainty-aware predictions (based on GP-PoM or indicator imputations) are beneficial for predictive performance, even compared to the uncertainty-aware training of conventional  ...  The signature transform is a 'universal nonlinearity' on the space of continuous vector-valued paths, and has received attention for use in machine learning on time series.  ...  Overall, we find that uncertainty-aware approaches (indicator imputation and GP-PoM) are beneficial when imputing irregularly-spaced time series for classification.  ... 
arXiv:2005.12359v2 fatcat:z2kcwx545zgd3jv2twb54r7flu

Multi-view Integration Learning for Irregularly-sampled Clinical Time Series [article]

Yurim Lee, Eunji Jun, Heung-Il Suk
2021 arXiv   pre-print
In this work, we propose a multi-view features integration learning from irregular multivariate time series data by self-attention mechanism in an imputation-free manner.  ...  Electronic health record (EHR) data is sparse and irregular as it is recorded at irregular time intervals, and different clinical variables are measured at each observation point.  ...  The classical machine learning approaches for handling the irregularly sampled time series are mostly based on convolutional neural network (CNN), recurrent neural network (RNN), and more recently attention-based  ... 
arXiv:2101.09986v2 fatcat:u46qtifgbjer3odmjtmu3ynxfq

Self-Supervised Transformer for Sparse and Irregularly Sampled Multivariate Clinical Time-Series [article]

Sindhu Tipirneni, Chandan K. Reddy
2022 arXiv   pre-print
Experiments on real-world multivariate clinical time-series benchmark datasets demonstrate that STraTS has better prediction performance than state-of-the-art methods for mortality prediction, especially  ...  To tackle this problem, we propose a Self-supervised Transformer for Time-Series (STraTS) model which overcomes these pitfalls by treating time-series as a set of observation triplets instead of using  ...  ACKNOWLEDGMENTS We thank Lakshmi Tipirneni for her help with the clinical domain knowledge related to the extraction of our time-series dataset from MIMIC-III database and for providing clinical insights  ... 
arXiv:2107.14293v2 fatcat:bzcfmzb55jcxjb2rnaj2kbrfji

Predictive Analytics for Caring and Managing Acute Disease Patients: A Deep Learning–Based Method to Predict Crucial Complications Phenotypes (Preprint)

Jessica Qiuhua Sheng, Paul Jen-Hwa Hu, Xiao Liu, Ting-Shuo Huang, Yu-Hsien Chen
2020 Journal of Medical Internet Research  
insufficiencies and account for variations in phenotypic expressions.  ...  To predict crucial complication phenotypes among patients with acute diseases, we propose a novel, deep learning-based method that uses recurrent neural network-based sequence embedding to represent disease  ...  Acknowledgments The authors thank the Research Institute of Chang Gung Memorial Hospital for their assistance in data collection and preprocessing.  ... 
doi:10.2196/18372 pmid:33576744 fatcat:atowfq2ylnhu5alcidkqdhnmva

Learning to Detect Sepsis with a Multitask Gaussian Process RNN Classifier [article]

Joseph Futoma, Sanjay Hariharan, Katherine Heller
2017 arXiv   pre-print
Our proposed framework models the multivariate trajectories of continuous-valued physiological time series using multitask Gaussian processes, seamlessly accounting for the high uncertainty, frequent missingness  ...  The Gaussian process is directly connected to a black-box classifier that predicts whether a patient will become septic, chosen in our case to be a recurrent neural network to account for the extreme variability  ...  This project was partially funded by the Duke Institute for Health Innovation.  ... 
arXiv:1706.04152v1 fatcat:lzyiwzcp4fgthm2dabansbiium

Quality control, data cleaning, imputation [article]

Dawei Liu, Hanne I. Oberman, Johanna Muñoz, Jeroen Hoogland, Thomas P.A. Debray
2021 arXiv   pre-print
A gentle introduction is provided on common statistical and machine learning methods for imputation.  ...  Finally, we introduce alternative methods to address incomplete data without the need for imputation.  ...  A first type of RNN methods generate multiple imputed datasets, and include Bidirectional Recurrent Imputation for Time Series (117) , multi-directional recurrent neural networks (118) , and residual  ... 
arXiv:2110.15877v1 fatcat:tzswjt4kerd7zegbgt5mdm23gi

Temporal Clustering with External Memory Network for Disease Progression Modeling [article]

Zicong Zhang, Changchang Yin, Ping Zhang
2021 arXiv   pre-print
To address these two issues, we propose Temporal Clustering with External Memory Network (TC-EMNet) for DPM that groups patients with similar trajectories to form disease clusters/stages.  ...  for producing comprehensive patient states.  ...  [22] proposed to use VAE to impute missing values for electronic health data with uncertainty-aware attention.  ... 
arXiv:2109.14147v2 fatcat:tjm4eqplnvaapfel44dbv72lme

Probabilistic Machine Learning for Healthcare [article]

Irene Y. Chen, Shalmali Joshi, Marzyeh Ghassemi, Rajesh Ranganath
2020 arXiv   pre-print
Beyond predictive models, we also investigate the utility of probabilistic machine learning models in phenotyping, in generative models for clinical use cases, and in reinforcement learning.  ...  Krishnan, Peter Schulam, and Pete Szolovits for helpful and useful feedback. This work was supported in part by a CIFAR AI Chair at the Vector Institute (MG) and Microsoft Research (MG).  ...  Using these indicators of observation in a time series, recurrent neural networks have been used to predict patient outcomes in the intensive care unit (11) .  ... 
arXiv:2009.11087v1 fatcat:htosfeqvhndvfmlmud2pvl3nsy

AttDMM: An Attentive Deep Markov Model for Risk Scoring in Intensive Care Units [article]

Yilmazcan Özyurt, Mathias Kraus, Tobias Hatt, Stefan Feuerriegel
2021 arXiv   pre-print
In this work, we propose a novel generative deep probabilistic model for real-time risk scoring in ICUs. Specifically, we develop an attentive deep Markov model called AttDMM.  ...  Clinical practice in intensive care units (ICUs) requires early warnings when a patient's condition is about to deteriorate so that preventive measures can be undertaken.  ...  70.04 Systolic blood pressure time-series 49.36 Heart rate time-series 48.69 Body temperature time-series 70.46 PaO 2 /FiO 2 time-series 93.88 Urinary output time-series 57.16 Serum urea  ... 
arXiv:2102.04702v2 fatcat:3rhtfxfocnavvkipwlse5x2gbm

Scalable Joint Models for Reliable Uncertainty-Aware Event Prediction [article]

Hossein Soleimani, James Hensman, Suchi Saria
2017 arXiv   pre-print
Imputation methods, which are typically used for completing the data prior to event prediction, lack a principled mechanism to account for the uncertainty due to missingness.  ...  Missing data and noisy observations pose significant challenges for reliably predicting events from irregularly sampled multivariate time series (longitudinal) data.  ...  ACKNOWLEDGMENTS The authors thank Wenbo Pan for his help with an initial implementation of the algorithm and Katharine Henry and Andong Zhang for their help with data.  ... 
arXiv:1708.04757v1 fatcat:3luywvmi5jaejbph4c34z526ey

A Multi-Agent Approach for Personalized Hypertension Risk Prediction

Sundus Abrar, Chu Kiong Loo, Naoyuki Kubota
2021 IEEE Access  
The time series is divided into time bins corresponding to the total number of seasonal variations that can be expected in the time series.  ...  Therefore, we utilize the same approach to model this seasonal variation in the blood pressure data and detect any outliers in the time series.  ... 
doi:10.1109/access.2021.3074791 fatcat:wanpehawkbgutge5aunktmwzcu

Systematic Review of Using Machine Learning in Imputing Missing Values

Mustafa Alabadla, Fatimah Sidi, Iskandar Ishak, Hamidah Ibrahim, Lilly Suriani Affendey, Zafienas Che Ani, Marzanah A. Jabar, Umar Ali Bukar, Navin Kumar Devaraj, Ahmad Sobri Muda, Anas Tharek, Noritah Omar (+1 others)
2022 IEEE Access  
The authors would like to thank the anonymous reviewers for their comments.  ...  For instance, recurrent neural networks are usually utilized to address missing values in longitudinal or time-series datasets [108] .  ...  Different datasets have different data types, such as time-series datasets, which work better with recurrent neural networks [108] .  ... 
doi:10.1109/access.2022.3160841 fatcat:uxrqlpsdaretrhbnccibqwzvi4

SurvLatent ODE : A Neural ODE based time-to-event model with competing risks for longitudinal data improves cancer-associated Deep Vein Thrombosis (DVT) prediction [article]

Intae Moon, Stefan Groha, Alexander Gusev
2022 arXiv   pre-print
To that end, we propose a generative time-to-event model, SurvLatent ODE, which adopts an Ordinary Differential Equation-based Recurrent Neural Networks (ODE-RNN) as an encoder to effectively parameterize  ...  SurvLatent ODE outperforms the current clinical standard Khorana Risk scores for stratifying DVT risk groups.  ...  Mayer Foundation, DDCF Clinical Scientist Development Award, Phi Beta Psi Sorority, and The Emerson Collective.  ... 
arXiv:2204.09633v1 fatcat:blc3ubyjvnb25mnhylacv35vey

Anomaly Detection Framework for Wearables Data: A Perspective Review on Data Concepts, Data Analysis Algorithms and Prospects

Jithin S. Sunny, C. Pawan K. Patro, Khushi Karnani, Sandeep C. Pingle, Feng Lin, Misa Anekoji, Lawrence D. Jones, Santosh Kesari, Shashaanka Ashili
2022 Sensors  
Clinical studies employing wearable devices are also increasing. In this article, we review the nature of the wearables-associated data and the downstream processing methods for detecting anomalies.  ...  Given the clinical significance of anomalies and their impact on diagnosis and treatment, a wide range of detection methods have been proposed to detect anomalies.  ...  Acknowledgments: The authors would like to thank Manasa Tata and Omnia Heikal from Rhenix Lifesciences for their contributions in reviewing the manuscript and preparing the figures, respectively.  ... 
doi:10.3390/s22030756 pmid:35161502 pmcid:PMC8840097 fatcat:nlju2m7btjccxk6xz6tfl6jyt4
« Previous Showing results 1 — 15 out of 1,372 results