​Master’s Thesis at the University of Basrah Explores Hybridizing Kriging with Adaptive LASSO

A master’s thesis at the College of Administration and Economics, University of Basrah, has explored the hybridization of a multi-task Kriging model with adaptive LASSO regression to improve the accuracy of spatial estimation.

The thesis, presented by student Abdullah Salem Kasib, aims to propose a hybrid statistical model that enhances the efficiency and stability of spatial prediction for high-dimensional, multi-task data, outperforming traditional models.

The research covers the general framework of Kriging models, variable selection techniques, and the mechanism for hybridizing them using the Least Angle Regression (LARS) algorithm.

The study concludes that the hybrid model is superior, demonstrating greater numerical efficiency in both simulations and practical applications—specifically regarding salinity levels in the Shatt al-Arab.

The thesis recommends the adoption of the proposed models and suggests incorporating a larger set of variables in future research.