Intelligibility of Nonparametric Survival Analysis for Health Security Policy Evaluation: Application to the Analysis of COVID-19 Data

Intelligibility of Nonparametric Survival Analysis for Health Security Policy Evaluation: Application to the Analysis of COVID-19 Data

Hamlili Ali
Copyright: © 2022 |Pages: 40
DOI: 10.4018/978-1-6684-2304-2.ch006
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Survival analysis is one of the most important research topics in probability and statistics applications to health and medical data. Its implementation has caught the attention of a large community of researchers from several skills including data analysis, statistical modeling, data mining, data science, and artificial intelligence. Survivor function and death intensity allow the analyst to assess the dynamics of the proportion of deaths and the risk of death over time. This chapter proposes an approach to the analysis of interval-censored survival data based on plug-in estimation as well as tools to assess the predictive quality of these estimators. Graphical tests are provided to support the choice between health security policies with the objective of leading to low mortality. The intertwining of this method with artificial intelligence tools paves the way for both personalized, precise, and efficient medicine and automated, informed, and rapid decision-making in health security.
Chapter Preview
Top

Introduction

To meet the needs of collecting and interpreting medical and health data, it is necessary more than ever to introduce methods of machine learning (ML) and artificial intelligence (AI). This will allow medicine to enter a new era where diagnoses will be more precise, clinical dashboards more complete and treatments more personalized. Also, it will allow health services and governments to fully exploit health data, share information, evaluate health security policies, improve them and fight effectively together against serious diseases and pandemics. The volume of data thus collected and shared could reach very large sizes. This is what, alongside the particularity (incomplete, censored, truncated) of this data, would require the intervention of very advanced methods of analysis. Survival analysis constitutes one of ML methods to meet this kind of need. It helps to answer questions that traditional data analysis methods are unable to resolve.

Complete Chapter List

Search this Book:
Reset