Relative Estimate of Revised Cardiovascular Risk Combining Traditional and Non-traditional Image-based CV Markers: A Kerala Based Study

Paul, Paulin; George, Noel

doi:10.2174/1573405616666200218125539

Cited by 3 publications

(3 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this proposed study, an optimized hybrid ASCV risk prediction and classification framework were studied to identify the 'present' and '10-years' ASCV risk status and risk class of a subject at an early stage. The input for the study has used JBS3 based Traditional Risk Factors (TRFs) [24] and ultrasound-based non-Traditional Risk Factors (non-TRFs) [25] that define the atherosclerotic risk status of a subject.…”

Section: Methodsmentioning

confidence: 99%

“…These derived risk scores describe the atherosclerotic risk status of a subject. The consolidated cIMT and cP risk scoring used in this study are based on the calculation described in the study [25]. Each column in the dataset is a specific CV risk indicator, and each row corresponds to the CV risk recordings of a subject.…”

Section: Data Descriptionmentioning

confidence: 99%

“…The ML-based framework implemented in this study was compared with similar studies in the literature, the established ASCV risk assessment tool, and the results from the statistical modelling as well. The ACC/AHA ASCVD risk prediction tool has demonstrated wide variations in performance when studied in different ethnicities [24,25,38]. Several studies have tried the combination of non-traditional US carotid markers of cIMT and cP burden status as a powerful link to indicate and predict atherosclerotic CV (ASCV) risk.…”

Section: Proposed Model Validity Against the Literaturementioning

confidence: 99%

See 2 more Smart Citations

GWO based Atherosclerotic Cardiovascular Prediction model integrating Traditional and Non-Traditional parameters

Paul

2022

Preprint

View full text Add to dashboard Cite

The study has proposed an optimized intelligent attribute reduction, risk prediction, and classification framework using Machine Learning for early identification of Atherosclerotic Cardiovascular (ASCV) risk status. Data preprocessing with normalization and missing data imputation using Multiple Imputation (MI) with Gaussian Copula (GC) method was done before optimal attribute selection. The proposed framework represented as SVMGWO-ANFISGWO-Multi-SVM has used Grey Wolf Optimization (GWO) algorithm. Feature selection was based on fitness evaluated wrapper-based multi-Support Vector Machine (multi-SVM). The optimal features selected by SVMGWO are processed by ANFISGWO for optimal ASCV risk prediction. Finally, multi-SVM is used to identify the risk class of the predicted ASCV risk. Performance of the study model were evaluated using Accuracy, Sensitivity, Specificity, Precision, G-mean, and F-measure features. Confusion and ROC plots are used to identify the classier performances. The Area Under Curve (AUC) was plotted to study the discrimination between the ASCV risk classes. The implementation was done in MATLAB 2015b. The ANFISGWO in the training phase has shown 80.7% accuracy, 100% sensitivity, 77.4% specificity, 43.3% precision, 60.2% G-mean, and 87.9% F-measure. Whereas, in the testing phase, the accuracy was 80%, 100% sensitivity, 77.1% specificity, 38.1% precision, 54.4% G-mean, and 87.7% F-measure. The overall AUC of 0.976 shows good discrimination between the ASCV risk classes.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Data Descriptionmentioning

confidence: 99%