Using Stacking methods based Genetic Algorithm to predict the time between symptom onset and hospital arrival in stroke patients and its related factors

Abstract

Introduction: Early arrival of patients with acute ischemic stroke to start of treatment by recombinant tissue plasminogen activator (rt-PA) within 4.5 hours after onset of stroke. We aimed to develop a machine learning method to predict effective factors on arrival time of patients with stroke to hospital after symptom onset. Methods: We included 676 patients with ischemic stroke who referred to Ardabil city hospital a province in northwest of Iran at year 2018. Classification models such as Random forest (RF), Gradient Boosting Classifier (GB), Decision Tree Classifier (DT), Support-Vector Machines (SVM), Naïve Bayes (NB), artificial neural networks (ANN), and logistic regression (LR) with 10-fold cross-validation were developed to predict effective factors on arrival time of patient with stroke to hospital. The performances were evaluated with accuracy, sensitivity, specificity, positive prophetical worth, and negative prophetical worth. Results: Of all patients, 25.3% arrived to the hospital in less than 4.5 hours. The accuracy of RF, NB, ANN, GB, DT, SVM, LR and suggest method (Stacking) were 0.98, 0.72, 0.73, 0.79, 0.98, 0.73, 0.74, and 0.99. Conclusion: In this study, the Stacking technique provide a better result (Accuracy 99.51%, sensitivity 100%, and specificity 99.40%) among all other techniques and this model could be used as a valuable tool for clinical decision making.

Keywords

• Stroke
• Machine learning
• Classification
• Random Forest
• Hospital