A Predictive Model for Guillain-Barré Syndrome Based on Single Learning Algorithms

Abstract

Background. Guillain-Barré Syndrome (GBS) is a potentially fatal autoimmune neurological disorder. The severity varies among the four main subtypes, named as Acute Inflammatory Demyelinating Polyneuropathy (AIDP), Acute Motor Axonal Neuropathy (AMAN), Acute Motor Sensory Axonal Neuropathy (AMSAN), and Miller-Fisher Syndrome (MF). A proper subtype identification may help to promptly carry out adequate treatment in patients. Method. We perform experiments with 15 single classifiers in two scenarios: four subtypes' classification and One versus All (OvA) classification. We used a dataset with the 16 relevant features identified in a previous phase. Performance evaluation is made by 10-fold cross validation (10-FCV). Typical classification performance measures are used. A statistical test is conducted in order to identify the top five classifiers for each case. Results. In four GBS subtypes' classification, half of the classifiers investigated in this study obtained an average accuracy above 0.90. In OvA classification, the two subtypes with the largest number of instances resulted in the best classification results. Conclusions. This study represents a comprehensive effort on creating a predictive model for Guillain-Barré Syndrome subtypes. Also, the analysis performed in this work provides insight about the best single classifiers for each classification case.

A Predictive Model for Guillain–Barré Syndrome Based on Ensemble Methods

“…We compare results from these ensembles against those obtained by single classifiers from a previous study [9]. Previously, we showed promising ensemble results by combining trees using C5 and Random Forest only [13].…”

“…The resulting dataset contains 156 variables: 121 variables from the nerve conduction test, four variables from the CSF analysis, and 31 clinical variables.We proposed a novel method combining quenching simulated annealing [9] (QSA) and Partitions around medoids (PAM) (the QSA-PAM method). QSA is a metaheuristic that generates approximate-to-the-optimal solutions in reasonable times for complex combinatorial problems.…”

“…We proposed a novel method combining quenching simulated annealing [9] (QSA) and Partitions around medoids (PAM) (the QSA-PAM method). QSA is a metaheuristic that generates approximate-to-the-optimal solutions in reasonable times for complex combinatorial problems.…”

“…SVMLap is not implemented in Weka 3.6.12 [14]. Therefore, we used SVMPoly (second best) [9] instead of AIDP versus AMSAN classification.…”

“…These five methods were applied in three test scenarios, four GBS subtypes classification, One versus One (OVO) classification, and One versus All (OVA) classification, and compared their performance. In a previous study [9], we investigated the performance of 15 different-in-nature classifiers such as decision trees (C4.5), instance-based learners ( k NN: k Nearest neighbor), kernel-based (SVM: Support vector machines), neural networks (SLP, MLP, RBF-DDA), and rule induction learners (OneR, JRip), among others. In this work, we made a performance comparison between single against ensemble classifiers.…”