Cubic-Regression and Likelihood Based Boosting GAM to Model Drug Sensitivity for Glioblastoma

Generalized additive models (GAM) being versatile and efficient do present some difficulties in model selection and inference based on context and fitting with linear smoothers. GAM is being extended to embed smooth functions of variables described by a parameter which governs curve smoothness or approximate forecast precision in a wide range of non‐linear models. Boosted GAM offers smooth function estimates of covariate influence functions and estimated degrees of freedom. GAM boosting solves the problem of limiting explanatory elements to relatively few and choosing smoothing settings. The challenges of drug Sensitivity towards the Glioblastoma are modelled for tissue samples, using the proposed Cubic‐Regression‐based GAM, and further analyzed using boosted GAM approach. Various spline variations and smooth interaction functions have been used to examine the behavior of the GAM model. Similarly, different base‐learners have been explored in boosted GAM, and a hyperparameter is tuned to analyze the model behavior.