I have two suggestions for possible improvements as follows:
Sparse matrix: In some practical problems (e.g., document classification) where the covariate matrix is extremely sparse, can skscope utilize the characteristic of sparse matrix to imporve the efficiency of storage and computation? Some reference may be helpful such as scipy.sparse or its JAX version jax.scipy.sparse.
Optimizers: I find that the base numeric solver for skscope is mainly based on nlopt now, and I wonder whether it is possible to support more optimizers. For example, the differentiable optimizer JAXopt may be more suitable for general (unconstrained, constrained and composite) optimization problems .
I have two suggestions for possible improvements as follows:
skscope
utilize the characteristic of sparse matrix to imporve the efficiency of storage and computation? Some reference may be helpful such as scipy.sparse or its JAX version jax.scipy.sparse.skscope
is mainly based onnlopt
now, and I wonder whether it is possible to support more optimizers. For example, the differentiable optimizer JAXopt may be more suitable for general (unconstrained, constrained and composite) optimization problems .Thanks!