Abstract

BackgroundCarotid plaque, recognized as a predictor of adverse cardiovascular and cerebrovascular complications, has been explored via machine learning. With Artificial intelligence in general and machine learning in particular exhibiting superior predictive and prognosticative capabilities, when compared with conventional statistical techniques, we explored the current state-of-the-art for automated Machine Learning to develop prediction algorithmic models for Carotid Plaque (CP) formation with the aim to optimise risk stratification and subsequent complication triaging among general population. MethodologyThe study population comprised 122 entrants— variables including subjective measures such as weight, height and smoking history as well as serum levels of HbA1c and HDL-C. The current state of the art (SOTA) for automated Machine Learning (aML) was adopted to develop predictive models using algorithms including Neural Network, eXtreme Gradient Boosting and CatBoost with the employment of hyperparameter tuning. Ensemble approach, which is the amalgamation of two or more than two algorithmic models to develop such a model which is better than either of its computive components, was superimposed. The training was achieved on 40% of the loaded dataset with the theretofore unexposed 60% of the said dataset used for testing that respective developed algorithmic model which came out to exhibit the highest macro weighted average Area Under the Receiver Operating Curve (mWA-AUROC) among the trained models. mWA-AUROC and accuracy, along with other parameters, were adopted to assess the predictive discriminative ability of the tested algorithmic model. ResultsA Decision Tree algorithmic model predicted CP formation with a perfect mWA-AUROC of 1.00 and an accuracy of 100% on the training partition (40% of the original dataset) along with an mWA-AUROC of 0.99 and an accuracy of 98.6% on the testing partition (60% of the original dataset). (Figure 1) A precision of 100% and a recall of 98.5%— such that an F1 score of 99.2%— was achieved on the testing partition by the respective algorithmic model— to be heretofore termed as the “Yurf-Bangash-Yousafzai augur”. The Yurf-Bangash-Yousafzai augur thus outsmarts the CP predictive model (having AUROC, accuracy and an F1 score each of 0.86) developed by Wu D et al. ConclusionsOur novel approach to developing a predictive risk model for CP formation by exploring SOTA for automated machine learning provides optimised predictions which, when incorporated into the respective risk stratification protocols, shall translate into a decrease in the morbidity and mortality associated with cardiovascular atherosclerotic events by assisting in risk stratification and subsequent complication triaging among the general population.These authors contributed equally to data curation, development of methodology, statistical analyses and writing with subsequent reviewing and editing of the abstract draft.These authors, indicated alphabetically, contributed equally to statistical analyses and validation.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call