Open AccessIntelligent Stroke Subtyping Using Recursive Elimination
Author(s) -
G. Lavanya,
S. Pradeep,
J. Udaya Prakash,
Susanne Prince
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1916/1/012078
Subject(s) - subtyping , random forest , classifier (uml) , computer science , feature selection , machine learning , artificial intelligence , adaboost , naive bayes classifier , ischemic stroke , data mining , pattern recognition (psychology) , support vector machine , medicine , ischemia , cardiology , programming language
Ischemic stroke subtyping is essential for the forecast of ischemic stroke apart from its usage in effective design and treatment of the same. The manual assessment of affliction grouping procedure is time-consuming, having limitation on dataset and is prone to error. This work considers feature selection and forecast problems in medical datasets. Shapiro-Wilk algorithm has been used to rank the features and Pearson correlations between features have been analyzed. Additionally, the proposed work uses the Recursive Feature Elimination with Cross-Validation (RFECV) using linear SVC, Random-Forest-Classifier, Extra-Trees-Classifier, AdaBoost-Classifier and multinomial - Naïve-Bayes-Classifier to select the important features. Then a simple deep learning model has been exploited to classify the ischemic stroke subtype on the International Stroke Trial (IST) dataset. The proposed method classifies the ischemic stroke subtype exactly and the results also proved that the machine learning approach performed well than the human professionals.