Simultaneous Estimation and Dataset Selection for Transfer Learning in High Dimensions by a Non-convex Penalty

In this paper, we introduce a method for simultaneous parameter estimation and informative source dataset identification in high-dimensional transfer learning, leveraging the truncated norm penalty function.This integrated approach contrasts with conventional strategies that treat useful dataset selection and transfer learning as separate steps.To solve the resulting non-convex optimization problem, specifically under sparse linear regression and generalized low-rank trace regression models, we adopt the difference of convex (DC) programming with the alternating direction method of multipliers (ADMM) procedure.We theoretically justify the proposed algorithm from both statistical and computational perspectives.Numerical results are reported alongside to validate the theoretical assertions.An R package MHDTL is developed to implement the Dong