Erythropoiesis stimulating agent recommendation model using recurrent neural networks for patient with kidney failure with replacement therapy

Yun, Hae‐Ryong; Lee, Gyubok; Jeon, Myeong Jun; Kim, Hyung Woo; Joo, Young Su; Kim, Hyoungnae; Chang, Tae Ik; Park, Sun-Hee; Han, Seung Hyeok; Kang, Shin Wook; Kim, Wooju; Yoo, Tae‐Hyun

doi:10.1016/j.compbiomed.2021.104718

Cited by 7 publications

(7 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After decades of research, there is still no specific treatment for sepsis. The improvement in patient outcomes came primarily from nonspecific interventions, including fluid resuscitation, early application of antibiotics, and elimination of the source of infection ([ 3 ] #5; [ 4 ] #8478; [ 5 ] #8582; [ 6 ] #49). An important reason for this disheartening situation is that the definitions of sepsis and septic shock cover a very heterogeneous population of patients.…”

Section: Introductionmentioning

confidence: 99%

Identification of Nine mRNA Signatures for Sepsis Using Random Forest

Zhou

Dong²,

Wang

et al. 2022

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

Sepsis has high fatality rates. Early diagnosis could increase its curating rates. There were no reliable molecular biomarkers to distinguish between infected and uninfected patients currently, which limit the treatment of sepsis. To this end, we analyzed gene expression datasets from the GEO database to identify its mRNA signature. First, two gene expression datasets (GSE154918 and GSE131761) were downloaded to identify the differentially expressed genes (DEGs) using Limma package. Totally 384 common DEGs were found in three contrast groups. We found that as the condition worsens, more genes were under disorder condition. Then, random forest model was performed with expression matrix of all genes as feature and disease state as label. After which 279 genes were left. We further analyzed the functions of 279 important DEGs, and their potential biological roles mainly focused on neutrophil threshing, neutrophil activation involved in immune response, neutrophil-mediated immunity, RAGE receptor binding, long-chain fatty acid binding, specific granule, tertiary granule, and secretory granule lumen. Finally, the top nine mRNAs (MCEMP1, PSTPIP2, CD177, GCA, NDUFAF1, CLIC1, UFD1, SEPT9, and UBE2A) associated with sepsis were considered as signatures for distinguishing between sepsis and healthy controls. Based on 5-fold cross-validation and leave-one-out cross-validation, the nine mRNA signature showed very high AUC.

show abstract

Section: Introductionmentioning

confidence: 99%

Identification of Nine mRNA Signatures for Sepsis Using Random Forest

Zhou

Dong²,

Wang

et al. 2022

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

show abstract

“…The performance of the Hb prediction algorithms was evaluated using the mean square error (MSE), mean absolute error (MAE), and mean error (ME), commonly used indices in related studies [ 8 , 9 ].…”

Section: Resultsmentioning

confidence: 99%

“…Yun et al used gated recurrent unit (GRU) networks in an ESA recommendation system [ 9 ]. First, the historical data are used to indicate the Hb level at the future time point, and then the predicted Hb value is changed as the target Hb value.…”

Section: Introductionmentioning

confidence: 99%

Multi-head self-attention mechanism enabled individualized hemoglobin prediction and treatment recommendation systems in anemia management for hemodialysis patients

Yang¹,

Lee²,

Liao³

et al. 2023

Heliyon

View full text Add to dashboard Cite

“…Deep reinforcement learning (DRL) ( Wang et al, 2020 ) can perceive complex inputs and learn independent policies, as reinforcement learning has shown great potential in solving many problems ( Pan and Xu, 2016 ; Zhang et al, 2020 ; Yun et al, 2021 ; Kaur and Mittal, 2022 ; Raheb et al, 2022 ). DRL makes robotic arm control more smart, does not require accurate modeling of the environment, and can compensate for the shortcomings of traditional motion planning methods.…”

Section: Introductionmentioning

confidence: 99%

An enhanced deep deterministic policy gradient algorithm for intelligent control of robotic arms

Dong

Liu

et al. 2023

Front. Neuroinform.

View full text Add to dashboard Cite

Aiming at the poor robustness and adaptability of traditional control methods for different situations, the deep deterministic policy gradient (DDPG) algorithm is improved by designing a hybrid function that includes different rewards superimposed on each other. In addition, the experience replay mechanism of DDPG is also improved by combining priority sampling and uniform sampling to accelerate the DDPG’s convergence. Finally, it is verified in the simulation environment that the improved DDPG algorithm can achieve accurate control of the robot arm motion. The experimental results show that the improved DDPG algorithm can converge in a shorter time, and the average success rate in the robotic arm end-reaching task is as high as 91.27%. Compared with the original DDPG algorithm, it has more robust environmental adaptability.

show abstract

Erythropoiesis stimulating agent recommendation model using recurrent neural networks for patient with kidney failure with replacement therapy

Cited by 7 publications

References 29 publications

Identification of Nine mRNA Signatures for Sepsis Using Random Forest

Identification of Nine mRNA Signatures for Sepsis Using Random Forest

Multi-head self-attention mechanism enabled individualized hemoglobin prediction and treatment recommendation systems in anemia management for hemodialysis patients

An enhanced deep deterministic policy gradient algorithm for intelligent control of robotic arms

Contact Info

Product

Resources

About