Sparse RepresentationBased on the minimum number of coefficients, the energy of more signals is described as much as possible. The coefficient distribution of different types of signals under different transformations will also be different.
The significance of sparse representation
- The purpose of sparse expression is to reduce dimension and save space. After sparse expression, the dependencies between dimensions of feature vectors are reduced and they are more independent.
- The sparse constraints added when solving sparse expressions make the "basis" obtained after calculation have the same importance in explaining the data, with the goal of finding the explanatory factors hidden behind the data;
- The theory deals with sparse solutions to systems of linear equations and can be used to find relevant solutions with applications in image processing, signal processing, machine learning, medical imaging, and more.
Characteristics of sparse expression
- Feature Selection: Automatic feature selection
- Interpretability: Make the model easier to explain