Diagnosis and Prediction of Large-For-Gestational-Age Fetus Using the Stacked GeneralizationMethod

Abstract

An accurate and efficient Large-for-Gestational-Age (LGA) classification system isdeveloped to classify a fetus as LGA or non-LGA, which has the potential to assist paediatricians andexperts in establishing a state-of-the-art LGA prognosis process. The performance of the proposedscheme is validated by using LGA dataset collected from the National Pre-Pregnancy and ExaminationProgram of China (2010–2013). A master feature vector is created to establish primarily datapre-processing, which includes a features’ discretization process and the entertainment of missingvalues and data imbalance issues. A principal feature vector is formed using GridSearch-basedRecursive Feature Elimination with Cross-Validation (RFECV) + Information Gain (IG) featureselection scheme followed by stacking to select, rank, and extract significant features from the LGAdataset. Based on the proposed scheme, different features subset are identified and provided tofour different machine learning (ML) classifiers. The proposed GridSearch-based RFECV+IG featureselection scheme with stacking using SVM (linear kernel) best suits the said classification processfollowed by SVM (RBF kernel) and LR classifiers. The Decision Tree (DT) classifier is not suggestedbecause of its low performance. The highest prediction precision, recall, accuracy, Area Underthe Curve (AUC), specificity, and F1 scores of 0.92, 0.87, 0.92, 0.95, 0.95, and 0.89 are achievedwith SVM (linear kernel) classifier using top ten principal features subset, which is, in fact higherthan the baselines methods. Moreover, almost every classification scheme best performed with tenprincipal feature subsets. Therefore, the proposed scheme has the potential to establish an efficientLGA prognosis process using gestational parameters, which can assist paediatricians and experts toimprove the health of a newborn using computer aided-diagnostic system.