Comparative Analysis of ELM and Sparse Bayesian ELM for Healthcare Diagnosis

Abstract

Extreme Learning Machine is a popular technique that became a center for research because of its easier implementation. It is s faster machine learning algorithm that has shown to have higher accuracy in various classification problems and has proven to be less time-consuming than traditional neural networks. But it also suffers from various drawbacks which are resolved with the help of the Bayesian paradigm along with the use of a Bayesian system known as Automatic Relevance Determination. We have tested how it performs on 3 different kinds of datasets. Various metrics are computed for all the 3 datasets and compared with the results from the conventional ELM. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.