Output data

The SM-SPSA algorithm can also provide different types of output data. You can specify which data is provided and/or saved by setting several variables that can be found in smspsa/main.py. The “normal” and “memory optimised” versions of the algorithm provide different types of output data, although both can provide a running time table and run parameter overviews. The output data is discussed in separate sections.

Output data

• RUNNING_TIME_FILE_NAME
• RUN_PARAMETERS_FILE_NAME
• Output data SM-SPSA
  • Gradient estimates
  • Objective values
  • Stationary distributions
  • Matrices in the unconstrained space
  • Transition matrices
  • Matrix plots
  • Network plot
  • Objective plot
• Output data SM-SPSA memory optimised
  • Final gradient estimate
  • Objective values
  • Final stationary distribution
  • Final matrix in the unconstrained space
  • Final transition matrix
  • Network plot
  • Objective plot