DBN-Extended: A Dynamic Bayesian Network Model Extended With Temporal Abstractions for Coronary Heart Disease Prognosis

Orphanou, Kalia; Stassopoulou, Athena; Keravnou, Elpida

doi:10.1109/jbhi.2015.2420534

Cited by 38 publications

(14 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One approach to deal with data imbalance problems is through the use of resampling techniques [43]. In this work, we under-sampled the training and test sets from the majority class (i.e., non-cancer cases) to preserve a 1:1 ratio of the cancer to non-cancer cases.…”

Section: Evaluation and Resultsmentioning

confidence: 99%

“…But given the class imbalance present in the dataset (1:24 cancer to non-cancer cases), we would not gain insight into the models’ ability for the more important predictive classification of cancer. [43] used similar methods to deal with imbalance in their dataset, but instead chose to oversample the minority class until a 1:1 ratio was achieved in their training set. They also reported metrics such as precision, recall, and F-score to compare performance against imbalanced datasets.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

Petousis

Han

Aberle

et al. 2016

Artificial Intelligence in Medicine

View full text Add to dashboard Cite

Introduction Identifying high-risk lung cancer individuals at an early disease stage is the most effective way of improving survival. The landmark National Lung Screening Trial (NLST) demonstrated the utility of low-dose computed tomography (LDCT) imaging to reduce mortality (relative to x-ray screening). As a result of the NLST and other studies, imaging-based lung cancer screening programs are now being implemented. However, LDCT interpretation results in a high number of false positives. A set of dynamic Bayesian networks (DBN) were designed and evaluated to provide insight into how longitudinal data can be used to help inform lung cancer screening decisions. Methods The LDCT arm of the NLST dataset was used to build and explore five DBNs for high-risk individuals. Three of these DBNs were built using a backward construction process, and two using structure learning methods. All models employ demographic, smoking status, cancer history, family lung cancer history, exposure risk factors, comorbidities related to lung cancer, and LDCT screening outcome information. Given the uncertainty arising from lung cancer screening, a cancer state-space model based on lung cancer staging was utilized to characterize the cancer status of an individual over time. The models were evaluated on balanced training and test sets of cancer and non-cancer cases to deal with data imbalance and overfitting. Results Results were comparable to expert decisions. The average area under the curve (AUC) of the receiver operating characteristic (ROC) for the three intervention points of the NLST trial was higher than 0.75 for all models. Evaluation of the models on the complete LDCT arm of the NLST dataset (N = 25, 486) demonstrated satisfactory generalization. Consensus of predictions over similar cases is reported in concordance statistics between the models’ and the physicians’ predictions. The models’ predictive ability with respect to missing data was also evaluated with the sample of cases that missed the second screening exam of the trial (N = 417). The DBNs outperformed comparison models such as logistic regression and naïve Bayes. Conclusion The lung cancer screening DBNs demonstrated high discrimination and predictive power with the majority of cancer and non-cancer cases.

show abstract

Section: Evaluation and Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

Petousis

Han

Aberle

et al. 2016

Artificial Intelligence in Medicine

View full text Add to dashboard Cite

show abstract

“…Temporal Abstraction methods are thought to manage a switch from a qualitative time-stamped description of raw data to a qualitative interval-based representation of time series, with the main goal of abstracting high-level concepts from time-stamped data. In the literature, there are works approaching health processes with TA in some areas, such as the costs’ evaluation associated with Diabetes Mellitus [ 35 ], the prognosis of the risk for coronary heart disease [ 36 ], or for defining typical medial abstraction patterns [ 37 ]. These works tried to create an automatic summarising of patient’s current state based on patient’s data through temporal abstraction, nevertheless, most of the clinical variables (such as weight, blood pressure or blood glucose) have numerical values, and TA techniques are based on discrete labels, excluding important information from the analysis.…”

Section: Related Workmentioning

confidence: 99%

Dynamic Models Supporting Personalised Chronic Disease Management through Healthcare Sensors with Interactive Process Mining

Valero-Ramón

Fernández-Llatas

Valdivieso

2020

Sensors

View full text Add to dashboard Cite

Rich streams of continuous data are available through Smart Sensors representing a unique opportunity to develop and analyse risk models in healthcare and extract knowledge from data. There is a niche for developing new algorithms, and visualisation and decision support tools to assist health professionals in chronic disease management incorporating data generated through smart sensors in a more precise and personalised manner. However, current understanding of risk models relies on static snapshots of health variables or measures, rather than ongoing and dynamic feedback loops of behaviour, considering changes and different states of patients and diseases. The rationale of this work is to introduce a new method for discovering dynamic risk models for chronic diseases, based on patients’ dynamic behaviour provided by health sensors, using Process Mining techniques. Results show the viability of this method, three dynamic models have been discovered for the chronic diseases hypertension, obesity, and diabetes, based on the dynamic behaviour of metabolic risk factors associated. This information would support health professionals to translate a one-fits-all current approach to treatments and care, to a personalised medicine strategy, that fits treatments built on patients’ unique behaviour thanks to dynamic risk modelling taking advantage of the amount data generated by smart sensors.

show abstract

“…For the nominal type data in the dataset, literature [28,29] encode it with the digital coding method. The corresponding field is coded as 0, 1, ⋯ , according to the number of possible values for the field.…”

Section: Data Processingmentioning

confidence: 99%

Improved Deep Belief Networks (IDBN) Dynamic Model-Based Detection and Mitigation for Targeted Attacks on Heavy-Duty Robots

Xie

et al. 2018

Applied Sciences

View full text Add to dashboard Cite

Featured Application: The research is mainly used for heavy-duty robots to detect and mitigate target attacks, which come from both the cyber-domain and the physical-domain. Abstract:In recent years, the robots, especially heavy-duty robots, have become the hardest-hit areas for targeted attacks. These attacks come from both the cyber-domain and the physical-domain. In order to improve the security of heavy-duty robots, this paper proposes a detection and mitigation mechanism which based on improved deep belief networks (IDBN) and dynamic model. The detection mechanism consists of two parts: (1) IDBN security checks, which can detect targeted attacks from the cyber-domain; (2) Dynamic model and security detection, used to detect the targeted attacks which can possibly lead to a physical-domain damage. The mitigation mechanism was established on the base of the detection mechanism and could mitigate transient and discontinuous attacks. Moreover, a test platform was established to carry out the performance evaluation test for the proposed mechanism. The results show that, the detection accuracy for the attack of the cyber-domain of IDBN reaches 96.2%, and the detection accuracy for the attack of physical-domain control commands reaches 94%. The performance evaluation test has verified the reliability and high efficiency of the proposed detection and mitigation mechanism for heavy-duty robots.

show abstract

DBN-Extended: A Dynamic Bayesian Network Model Extended With Temporal Abstractions for Coronary Heart Disease Prognosis

Cited by 38 publications

References 16 publications

Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

Dynamic Models Supporting Personalised Chronic Disease Management through Healthcare Sensors with Interactive Process Mining

Improved Deep Belief Networks (IDBN) Dynamic Model-Based Detection and Mitigation for Targeted Attacks on Heavy-Duty Robots

Contact Info

Product

Resources

About