Feature Selection and Classification for High-Dimensional Incomplete Multimodal Data

Algorithm 3

Accelerated proximal gradient for solving problem (10).

Initialization: Initialize the Lipschitz constant

and set by warm start,

, , parameter , and

.

(1) Set .

(2) Set .

For ,

Set , compute .

if ,

 set , stop;

else

 .

End

end

(3) Set .

(4) Compute . Let

.

(5) Quit if stopping condition achieves. Otherwise, go

to step (1).