Contains code for computing empirically the mean-squared error of Lasso methods.

1. Algorithms for solving ell1 problems

iPython notebook that contains an AMP algorithm, and some other methods.

2. Empirical Lasso MSE

iPython notebook that uses spgl1 to compute the empirical MSE for several parameter settings in the underdetermined normal-random case.

3. jsonWriteDict

writes a dictionary containing numpy arrays to an easy-to-read json file.

4. pdmse

contains methods for computing the mean-squared error of proximal denoising solutions.

5. ProxDenoisSPGL1HistData

contains code that uses both pdmse and jsonWriteDict to compute the mean-squared error in the proximal denoising case a large number of times for several signal dimensions; stores the result in a json file.