Journal of Mathematical Problems, Equations and Statistics
  • Printed Journal
  • Refereed Journal
  • Peer Reviewed Journal

P-ISSN: 2709-9393, E-ISSN: 2709-9407

2023, Vol. 4, Issue 1, Part B


Stacked ensemble model for recurrent head and neck squamous cell carcinoma prognosis based on clinicopathologic and genomic markers


Author(s): Damianus Kofi Owusu and Peter Kwesi Nyarko

Abstract: The prevalence of head and neck squamous cell carcinoma (HNSCC) and its recurrences is not declining in Ghana as a result of the disease's delayed diagnosis and dismal prognosis. Early detection and treatment are crucial since HNSCC recurrence and tumor stage at diagnosis are significantly correlated. This study looked at the best meta-classifier model where the same ML classifiers for base classifiers and meta classifiers are employed in order to determine the most reliable prediction and robust prognostic model for recurrent HNSCC. Based on gradient boosted features (GBF), the suggested model was an ensemble of ML models that were stacked. Each of these models served as a meta-classifier and as a building block for the base classifiers. To find the optimal meta-classifier model, the performances of different meta-models were compared. The findings demonstrated that utilising the GBM as a meta-classifier produced superior accuracy with the least log loss compared to that produced by any other model of recurrent HNSCC prognostic data. This gave a stacked ensemble model termed as a HESCA model, consisted of five base models and GBM meta-model. 8-input HESCA model was compared with full-input model, and 8-input HESCA model was also compared with 8-input models. The results of the study demonstrated that using a GBM classifier as a meta-classifier in a stacking ensemble with five base classifiers based on GBF or GBM input features outperformed standalone models and any full-input model. Additionally, using a GBM as a meta-classifier is appropriate as a supporting tool for identifying, classifying, and predicting recurrent HNSCC prognosis data.

Pages: 121-134 | Views: 229 | Downloads: 73

Download Full Article: Click Here
How to cite this article:
Damianus Kofi Owusu and Peter Kwesi Nyarko. Stacked ensemble model for recurrent head and neck squamous cell carcinoma prognosis based on clinicopathologic and genomic markers. Journal of Mathematical Problems, Equations and Statistics. 2023; 4(1): 121-134.
Journal of Mathematical Problems, Equations and Statistics

Journal of Mathematical Problems, Equations and Statistics

Journal of Mathematical Problems, Equations and Statistics
Call for book chapter