Classification of Guillain–Barré Syndrome Subtypes Using Sampling Techniques with Binary Approach

Abstract

Guillain–Barré Syndrome (GBS) is an unusual disorder where the body’s immune system affects the peripheral nervous system. GBS has four main subtypes, whose treatments vary among them. Severe cases of GBS can be fatal. This work aimed to investigate whether balancing an original GBS dataset improves the predictive models created in a previous study. purpleBalancing a dataset is to pursue symmetry in the number of instances of each of the classes.The dataset includes 129 records of Mexican patients diagnosed with some subtype of GBS. We created 10 binary datasets from the original dataset. Then, we balanced these datasets using four different methods to undersample the majority class and one method to oversample the minority class. Finally, we used three classifiers with different approaches to creating predictive models. The results show that balancing the original dataset improves the previous predictive models. The goal of the predictive models is to identify the GBS subtypes applying Machine Learning algorithms. It is expected that specialists may use the model to have a complementary diagnostic using a reduced set of relevant features. Early identification of the subtype will allow starting with the appropriate treatment for patient recovery. This is a contribution to exploring the performance of balancing techniques with real data.

Diagnosis and treatment of Guillain-Barre using the prolog expert system

“…In contrast, the proposed solution stands out for its logical approach and explicit rules, prioritizing explainability and understanding in medical decision-making. Moreover, ML-based approaches, such as those presented by [13], [15], [18], yield solid results in terms of sensitivity, specificity and accuracy, demonstrating their effectiveness in the diagnostic task. On the other hand, the solution proposed in the present research in Prolog provides an essential nuance by focusing on the interpretability of the results, providing a deeper and reasoned understanding of diagnostic decisions.…”

“…In this comparative analysis, three distinct approaches to the diagnosis of GBS subtypes, presented by [13], [15], [18] and the proposed in this research for the diagnostic of GBS subtype, are evaluated, exploring the diverse perspectives that emerge in the convergence of medicine and artificial intelligence. While the four approaches share the fundamental objective of diagnosing GBS subtypes, they differ significantly in terms of evaluation methods, development tools, and results obtained.…”

“…This method will help medical professionals and lay the foundation for future research. Similarly, Torres-Vásquez et al [18], aimed to determine whether balancing an original GBS dataset would improve the ability of current predictive models to recognize GBS subtypes. They created ten binary datasets and used class balancing techniques on records of 129 Mexican patients with various GBS subtypes.…”

A Hybrid Sampling Approach for Improving the Classification of Imbalanced Data Using ROS and NCL Methods

“…It is widely recognized as an effective undersampling technique, known for its ability to remove data while preserving high quality. NCL prioritizes information cleaning and removing noise from the training data over balancing class proportions [54]. The undersampling process starts by identifying the N1 sample and its three nearest neighbors in the training data.…”

“…Conversely, if N1 is a part of the minority class, the majority group will be removed as its neighbor [55]. The steps of the NCL algorithm are outlined as follows [54].…”