Abstract
Artificial intelligence is an emerging area of modern research that aims at infusing machine intelligence through computational techniques. Precision Autism is the term coined to view Autism Spectrum Disorders (ASD) through a physiological lens than a purely psychological one. Recent research on ASD has revealed the unfavorable heterogeneous gamut to which this neurocognitive degeneration holds sway. Genes linked to autism in the SFARI database are also associated with genomic information that is linked to 25 other non-neurological diseases. Most ASD individuals were found to suffer from Obesity, gastrointestinal ailments, diabetes, hypertension, and other metabolic disorders. This research work aims at exploring the impact of conventional, deep, and Automated machine learning techniques in hepatic ailment classification as the initial investigation step before applying machine learning techniques on genetic data to unearth the relationship between ASD and hepatic ailments. This article places focus on two different datasets comprising of more than 900 patient records acquired from the University of California, Irvine, Machine Learning Repository. A predictive data mining framework that compares the performance of supervised learning classifiers in hepatic disorder classification based on Mathew's Correlation Coefficient and Balanced Accuracy is presented here. To the best of our knowledge, this is the first attempt to explore automated machine learning in liver disorder classification and to employ standard statistical measures for unbalanced data in rating classifier performance. Deep Neural Networks yielded the highest performance in classification with balanced accuracy ∼74% and an MCC of 0.46 on the BUPA data.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call