Orthogonal Discriminant Projection Dimension Reduction Method based on Firefly Algorithm Optimization

The traditional orthogonal discriminant projection dimension reduction method ignores the difference between categories in the dimension reduction process, and it is easy to fall into the local optimal solution when optimizing the projection vector, and it is sensitive to the selection of the initial projection vector. To solve this problem, this paper proposes an orthogonal discriminant projection dimension reduction method based on firefly algorithm optimization. By adjusting the brightness and attraction between fireflies to optimize the projection matrix, the discriminant information of the data can be better retained in the low-dimensional space, thereby improving the separability of the data.