Sparse Bayesian multitask learning for radar target recognition

A multitask learning framework is proposed to jointly model the tasks by sharing the subspace in this paper. Firstly, we assume the tasks (the classification of one class versus others) are related. Then the task predictors may belong to a low dimensional subspace to realize the information sharing. The subspace bases of each task predictor can be learned by utilizing the Bernoulli-Beta prior from the given data, and the number of shared bases between different tasks can be utilized to automatically determine their relatedness. The proposed method is validated on the measured radar data.