| CCA Algorithm |
|
|
|
|
| , | Given
a data set I of dimension M, CCA projects I to data set O of dimension S
where S<M |
||
| and
preserves in the output set O the distances between all the data points in
I |
|||
| Minimizing cost
function: |
|
|
| : a bounded and
monotonically |
|
| decreasing
function, e.g., exponential, etc |
|
| By updating in each
cycle: |
| … |
| … |
| , |
| Nx2 |
| NxN |
