Shandian Zhe’s research while affiliated with University of Utah and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (5)


Feature selection and hybrid machine learning (ML) model. The raw cardiothoracic (CT) surgery dataset contained numerous features. To reduce the features to a manageable number, a random forest feature selection procedure was performed. The selected features were then used as inputs into the Gaussian Process (GP) regression and classification ML algorithms.
Area under the curve (AUC) for each machine learning (ML) classification algorithm for the development phase. Here, we used only the initial dataset (n = 2410) with five-fold crossvalidation to reduce overfitting. Gaussian Process (GP) demonstrated the greatest AUC and overall performance.
Area under the curve (AUC) for each machine learning (ML) classification algorithm after the validation phase. Here, the initial dataset (n = 2410) was used as the training dataset and the additional cases included after updating the database (n = 437) were used as the test dataset. Gaussian Process (GP) again showed the highest AUC and best overall performance.
Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery
  • Article
  • Full-text available

January 2022

·

139 Reads

·

32 Citations

Zheng Wang

·

Shandian Zhe

·

·

[...]

·

Ryan A. Metcalf

Accurately predicting red blood cell (RBC) transfusion requirements in cardiothoracic (CT) surgery could improve blood inventory management and be used as a surrogate marker for assessing hemorrhage risk preoperatively. We developed a machine learning (ML) method to predict intraoperative RBC transfusions in CT surgery. A detailed database containing time-stamped clinical variables for all CT surgeries from 5/2014–6/2019 at a single center (n = 2410) was used for model development. After random forest feature selection, surviving features were inputs for ML algorithms using five-fold cross-validation. The dataset was updated with 437 additional cases from 8/2019–8/2020 for validation. We developed and validated a hybrid ML method given the skewed nature of the dataset. Our Gaussian Process (GP) regression ML algorithm accurately predicted RBC transfusion amounts of 0 and 1–3 units (root mean square error, RMSE 0.117 and 1.705, respectively) and our GP classification ML algorithm accurately predicted 4 + RBC units transfused (area under the curve, AUC = 0.826). The final prediction is the regression result if classification predicted < 4 units transfused, or the classification result if 4 + units were predicted. We developed and validated an ML method to accurately predict intraoperative RBC transfusions in CT surgery using local data.

Download

Macroscopic traffic flow modeling with physics regularized Gaussian process: A new insight into machine learning applications in transportation

April 2021

·

195 Reads

·

108 Citations

Transportation Research Part B Methodological

Despite the wide implementation of machine learning (ML) technique in traffic flow modeling recently, those data-driven approaches often fall short of accuracy in the cases with a small or noisy training dataset. To address this issue, this study presents a new modeling framework, named physics regularized machine learning (PRML), to encode classical traffic flow models (referred as physics models) into the ML architecture and to regularize the ML training process. More specifically, leveraging the Gaussian process (GP) as the base model, a stochastic physics regularized Gaussian process (PRGP) model is developed and a Bayesian inference algorithm is used to estimate the mean and kernel of the PRGP. A physics regularizer, based on macroscopic traffic flow models, is also developed to augment the estimation via a shadow GP and an enhanced latent force model is used to encode physical knowledge into the stochastic process. Based on the posterior regularization inference framework, an efficient stochastic optimization algorithm is then developed to maximize the evidence lowerbound of the system likelihood. For model evaluations, this paper conducts empirical studies on a real-world dataset which is collected from a stretch of I-15 freeway, Utah. Results show the new PRGP model can outperform the previous compatible methods, such as calibrated traffic flow models and pure machine learning methods, in estimation precision and is more robust to the noisy training dataset.


Graph constraint-based robust latent space low-rank and sparse subspace clustering

June 2020

·

410 Reads

·

6 Citations

Neural Computing and Applications

Recently, low-rank and sparse representation-based methods have achieved great success in subspace clustering, which aims to cluster data lying in a union of subspaces. However, most methods fail if the data samples are corrupted by noise and outliers. To solve this problem, we propose a novel robust method that uses the F-norm for dealing with universal noise and the l1l_1 norm or the l2,1l_{2,1} norm for capturing outliers. The proposed method can find a low-dimensional latent space and a low-rank and sparse representation simultaneously. To preserve the local manifold structure of the data, we have adopted a graph constraint in our model to obtain a discriminative latent space. Extensive experiments on several face benchmark datasets show that our proposed method performs better than state-of-the-art subspace clustering methods.


Macroscopic Traffic Flow Modeling with Physics Regularized Gaussian Process: A New Insight into Machine Learning Applications

February 2020

·

349 Reads

Despite the wide implementation of machine learning (ML) techniques in traffic flow modeling recently, those data-driven approaches often fall short of accuracy in the cases with a small or noisy dataset. To address this issue, this study presents a new modeling framework, named physics regularized machine learning (PRML), to encode classical traffic flow models (referred as physical models) into the ML architecture and to regularize the ML training process. More specifically, a stochastic physics regularized Gaussian process (PRGP) model is developed and a Bayesian inference algorithm is used to estimate the mean and kernel of the PRGP. A physical regularizer based on macroscopic traffic flow models is also developed to augment the estimation via a shadow GP and an enhanced latent force model is used to encode physical knowledge into stochastic processes. Based on the posterior regularization inference framework, an efficient stochastic optimization algorithm is also developed to maximize the evidence lowerbound of the system likelihood. To prove the effectiveness of the proposed model, this paper conducts empirical studies on a real-world dataset which is collected from a stretch of I-15 freeway, Utah. Results show the new PRGP model can outperform the previous compatible methods, such as calibrated pure physical models and pure machine learning methods, in estimation precision and input robustness.


Data-driven resource flexing for network functions visualization

July 2018

·

21 Reads

·

6 Citations

Resource flexing is the notion of allocating resources on-demand as workload changes. This is a key advantage of Virtualized Network Functions (VNFs) over their non-virtualized counterparts. However, it is difficult to balance the timeliness and resource efficiency when making resource flexing decisions due to unpredictable workloads and complex VNF processing logic. In this work, we propose an Elastic resource flexing system for Network functions VIrtualization (ENVI) that leverages a combination of VNF-level features and infrastructure-level features to construct a neural-network-based scaling decision engine for generating timely scaling decisions. To adapt to dynamic workloads, we design a window-based rewinding mechanism to update the neural network with emerging workload patterns and make accurate decisions in real time. Our experimental results for real VNFs (IDS Suricata and caching proxy Squid) using workloads generated based on real-world traces, show that ENVI provisions significantly fewer (up to 26%) resources without violating service level objectives, compared to commonly used rule-based scaling policies.

Citations (4)


... Some studies focus on the prediction of the required blood volume [6][7][8]. Most of these studies provide models for intra or perioperative prediction of BT for specific surgery types, like for spinal surgery [9], pelvic fracture patients [10], cardiac surgery [11,12], and gastric cancer [13]. There are only few studies targeting BT predictions from the perspective of an Intensive Care Unit (ICU) stay, which focus on gastrointestinal bleeding [14,15] or cardiothoracic surgery [11]. ...

Reference:

Predicting blood transfusion demand in intensive care patients after surgery by comparative analysis of temporally extended data selection
Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery

... PIGNN represents a subset of physics-informed machine learning (PINN) (Karniadakis et al., 2021), which embeds physics theories into machine learning to efficiently extract patterns from noisy data. The advantage of PINN has inspired many PINN transportation studies, including traffic flow (Yuan et al., 2021 and car-following modeling . However, despite the emergence of PIGNN studies in transportation (Zhu et al., 2022, Xue et al., 2024a, the optimal balance between physics-based and GNN components in PIGNN across various transportation contexts remains unclear. ...

Macroscopic traffic flow modeling with physics regularized Gaussian process: A new insight into machine learning applications in transportation
  • Citing Article
  • April 2021

Transportation Research Part B Methodological

... Besides, the computation cost of sparse and low-rank representations is very large when the dimension of the features is high [48]. To address these problems, some spatial projection-based approaches are proposed [48], [49], [50], [51], [52]. Motivated by them, we introduce projection matrices {W v } m v=1 ∈ R k×d v to project the original data into a lower-dimensional space for compact feature learning. ...

Graph constraint-based robust latent space low-rank and sparse subspace clustering

Neural Computing and Applications