Abstract:
The traditional regression model method is usually missing the label information. For this problem, this paper proposed a discriminant low-rank regression model method to optimization model. Firstly, we set the model objective matrix in advance, and improved the loss function using the local optimization and global optimization methods. Then, the augmented Lagrangian method was used to solve the objective function. A new model objective matrix was obtained on the basis of solving the function from the previous step, and the final mapping matrix was calculated through the linear regression model. Finally, the effectiveness of the proposed method was verified through the comparative experiments. The experimental results showed that compared with several other low-rank regression model methods, the proposed algorithm had higher recognition rat.