High-Risk Prediction of Cardiovascular Diseases via Attention-Based Deep Neural Networks

An, Ying; Huang, Nengjun; Chen, Xianlai; Wu, Fang‐Xiang; Wang, Jianxin

doi:10.1109/tcbb.2019.2935059

Cited by 46 publications

(19 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, the more features of miRNAs also should be designed, such as topological features of miR-NAs. Finally, other similarity-based methods [74], collaborative metric learning methods [75] and deep learning methods [76,77] should be adopted. We would provide a more effective computational method to predict essential miRNAs by addressing above limitations in the future.…”

Section: Resultsmentioning

confidence: 99%

PESM: predicting the essentiality of miRNAs based on gradient boosting machines and sequences

Cheng

Wang

et al. 2020

BMC Bioinformatics

Self Cite

View full text Add to dashboard Cite

Background: MicroRNAs (miRNAs) are a kind of small noncoding RNA molecules that are direct posttranscriptional regulations of mRNA targets. Studies have indicated that miRNAs play key roles in complex diseases by taking part in many biological processes, such as cell growth, cell death and so on. Therefore, in order to improve the effectiveness of disease diagnosis and treatment, it is appealing to develop advanced computational methods for predicting the essentiality of miRNAs. Result: In this study, we propose a method (PESM) to predict the miRNA essentiality based on gradient boosting machines and miRNA sequences. First, PESM extracts the sequence and structural features of miRNAs. Then it uses gradient boosting machines to predict the essentiality of miRNAs. We conduct the 5-fold cross-validation to assess the prediction performance of our method. The area under the receiver operating characteristic curve (AUC), F-measure and accuracy (ACC) are used as the metrics to evaluate the prediction performance. We also compare PESM with other three competing methods which include miES, Gaussian Naive Bayes and Support Vector Machine. Conclusion: The results of experiments show that PESM achieves the better prediction performance (AUC: 0.9117, F-measure: 0.8572, ACC: 0.8516) than other three computing methods. In addition, the relative importance of all features also further shows that newly added features can be helpful to improve the prediction performance of methods.

show abstract

Section: Resultsmentioning

confidence: 99%

PESM: predicting the essentiality of miRNAs based on gradient boosting machines and sequences

Cheng

Wang

et al. 2020

BMC Bioinformatics

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the medical domain, medical event prediction is a promising research topic. The main task of medical event prediction is to predict future medical events including risk of diseases [6], prescriptions [5], mortality rate [7], hospital readmission [8], length of stay in hospital [9], postoperative complications [10], survival time [11] and so on. In this paper, we mainly focus on next-period prescription prediction.…”

Section: Medical Event Predictionmentioning

confidence: 99%

“…Among these variants, the decomposed LSTM that used LSTM to model history prescription sequence and integrated laboratory test results into each cell achieved the best results. An et al proposed an attention-based LSTM for high-risk prediction [6]. The model first adopted attention-based LSTM to individually represent each type of medical information, including diagnosis sequence, laboratory sequence, and their combination sequence, and then concatenate these three representations together for prediction.…”

Section: Introductionmentioning

confidence: 99%

A hybrid method of recurrent neural network and graph neural network for next-period prescription prediction

Liu

Ding³

et al. 2020

Int. J. Mach. Learn. & Cyber.

View full text Add to dashboard Cite

Electronic health records (EHRs) have been widely used to help physicians to make decisions by predicting medical events such as diseases, prescriptions, outcomes, and so on. How to represent patient longitudinal medical data is the key to making these predictions. Recurrent neural network (RNN) is a popular model for patient longitudinal medical data representation from the view of patient status sequences, but it cannot represent complex interactions among different types of medical information, i.e., temporal medical event graphs, which can be represented by graph neural network (GNN). In this paper, we propose a hybrid method of RNN and GNN, called RGNN, for next-period prescription prediction from two views, where RNN is used to represent patient status sequences, and GNN is used to represent temporal medical event graphs. Experiments conducted on the public MIMIC-III ICU data show that the proposed method is effective for next-period prescription prediction, and RNN and GNN are mutually complementary.

show abstract

“…Based on the tuned image as shown in Figure.7. In particular, the constructed CDRAESM Framework by increasing cross-entropy loss with the aid of backpropagation algorithm which is solved in the following equation (15) and (16) The reconstructed error data for the feature j and i is represented as in the following equation 17& (18)…”

Section: Risk Information Is Preserved Using Proposed Cdraesm For mentioning

confidence: 99%

“…This paper aims to determine the capacity for an important amount of heterogeneous PER data; therefore, a new method of CAS prediction through deep learning in clinical risk [18] has been analyzed. The designed and developed Convolution Denoising Regularized Auto Encoder Stacked Method (CDRAESM) has particular benefits, including fast inference and the opportunity to recreate features with reasonable classification precision, from various deep learning models on IoMT platform.…”

Section: Introductionmentioning

confidence: 99%

Convolution Denoising Regularized Auto Encoder Stacked Method for Coronary Acute Syndrome in Internet of Medical Things Platform

et al. 2020

View full text Add to dashboard Cite

Presently, Coronary Acute Syndrome (CAS) is a widespread and extreme heart disease which is considered as one of the major concerns for death in the world with long-term disability. For early intervention and care, the prediction of CAS clinical risk is more critical during analysis. A minimum number of manually selected dimensions of risk are used for current CAS risk assessment models and statistical variables are often dichotomized to optimize storage in the Internet of medical things platform(IoMT) on the Encoder layer during data analysis. This research develops a Convolution Denoising Regularized Auto Encoder Stacked Method (CDRAESM) to normalize CAS patient's medical risks from high volumes of patient records and the characteristics have been analyzed during prediction. In this research, a true medical dataset of 3,464 CAS samples is used for experimental analysis and numerical reliability has been analyzed in Area Characteristics Curve (ACC) with an Accuracy range of 96.77%. The results show that the current health risk prediction using CDRAESM achieves competitive concert than conventional models which prevails in practice. Further, the reconstructive learning strategy approach can extract informational risk from the CAS and the risk factors has been identified with existing knowledge of the clinical domain and include theories that might be confirmed by further medical research. INDEX TERMS Coronary acute syndrome, convolution denoising, auto encoder stacked, area characteristics curve.

show abstract

High-Risk Prediction of Cardiovascular Diseases via Attention-Based Deep Neural Networks

Cited by 46 publications

References 42 publications

PESM: predicting the essentiality of miRNAs based on gradient boosting machines and sequences

PESM: predicting the essentiality of miRNAs based on gradient boosting machines and sequences

A hybrid method of recurrent neural network and graph neural network for next-period prescription prediction

Convolution Denoising Regularized Auto Encoder Stacked Method for Coronary Acute Syndrome in Internet of Medical Things Platform

Contact Info

Product

Resources

About