“Relative performance of model selection criteria based on Kullback’s Symmetric Divergence in survival analysis with applications using genetic data”
Country of Study
Benin
Institution
University of Abomey-Calavi
Expected Year of Completion
2024
Thematic Area
Life and Health Sciences
Education
Clarisse is pursuing her PhD in Agronomic and Water Sciences at the University of Abomey-Calavi and works as a Statistical Data Analyst at Green Tech Innovation. She has an MSc in Biostatistics from the same university and hopes to complete her studies in 2024.
Research Summary
Clarisse’s research investigates model selection criteria based on Kullback Symmetric Divergence in survival analysis. In statistical modelling, choosing the best model that characterises the underlying data within a collection of candidates is a crucial question. A biased model may significantly impact scientific interpretations as well as model predictions. For that, the model selection criteria are used. The model selection criteria are rules used to select a good statistical model among a set of candidate models. The most known are the Akaike Information Criterion (AIC). AIC is applicable in a broad array of modelling frameworks. Hurvich and Tsai (1989) suggested using the corrected AIC criterion when the sample size is small. These criteria have been developed using Kullback’s Directed Divergence. Other model selection criteria based on Kullback Symmetric Divergence (KSD) were developed. However, these last criteria have been found to outperform those based on Kullback’s directed divergence for linear regression, longitudinal data, and others. Unfortunately, neither KSD family is used to survival analysis. However, survival analysis has emerged as one of the most popular data analysis techniques in various fields, including medicine, criminology, marketing, astronomy, epidemiology, and environmental health, in the last four decades. Clarisse’s research will enable statisticians using survival analysis models to choose the optimal model based on the best criterion. Better prediction and decision-making will be possible as a result.
Publications:
1. Inventory of biosecurity measures and antibiotics therapy practices on laying hen farms in Benin
