Bleak Medical Prognosis of Covid Patients Using Explainable Machine Learning

Richa Sharma, Himanshu Pandey, Ambuj Kumar Agarwal, Dolley Srivastava

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Today, artificial intelligence (AI) algorithms are used to generate predictions in many high-stakes fields, such as credit risk analysis and medical diagnosis. Therefore, AI systems are having a greater impact on people, yet many cutting-edge technologies are opaque, negating people's 'right to explanation.' To address this issue, experts in the field have created explainable AI to detail the rationale behind an algorithm's prediction. The focus of this article is on creating an explanation system for a predicted outcome and a machine learning prediction model for poor medical prognosis in Covid patients. Many different types of machine learning algorithms, including both foundational and augmented versions, have been implemented. It was shown that the XGboost model provided superior diagnostic performance. This situation of a prediction has been explained using Shapley Additive Explanations (SHAP). The reason the XGBoost machine learning model gives a certain prediction result may be deduced with the use of three statistics charts.

Original languageEnglish
Title of host publicationProceedings of 3rd International Conference on Advanced Computing Technologies and Applications, ICACTA 2023
PublisherInstitute of Electrical and Electronics Engineers
ISBN (Electronic)9798350348347
DOIs
StatePublished - 1 Jan 2023
Externally publishedYes
Event3rd International Conference on Advanced Computing Technologies and Applications, ICACTA 2023 - Mumbai, India
Duration: 6 Oct 20237 Oct 2023

Publication series

NameProceedings of 3rd International Conference on Advanced Computing Technologies and Applications, ICACTA 2023

Conference

Conference3rd International Conference on Advanced Computing Technologies and Applications, ICACTA 2023
Country/TerritoryIndia
CityMumbai
Period6/10/237/10/23

Keywords

  • Covid-19
  • Explainable systems
  • Machine learning
  • SHAP value interpretation
  • XAI
  • XGBoost

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Computer Vision and Pattern Recognition
  • Control and Optimization

Fingerprint

Dive into the research topics of 'Bleak Medical Prognosis of Covid Patients Using Explainable Machine Learning'. Together they form a unique fingerprint.

Cite this