From 8df0ca7b88ba7958f014938e563ef480c0b1cebe Mon Sep 17 00:00:00 2001 From: Alex Date: Mon, 9 Dec 2024 08:44:39 +0000 Subject: [PATCH] Create README.md --- README.md | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) create mode 100644 README.md diff --git a/README.md b/README.md new file mode 100644 index 0000000..ffcaa9e --- /dev/null +++ b/README.md @@ -0,0 +1,25 @@ +Fit a spectral neural mass model, specified using the DCM conventional structure (DCM with DCM.M.pE, DCM.M.pC,DCM.M.f, DCM.M.IS and DCM.xY specified). + +Parameter estimation using log-likelihood (and/or Free Energy) optimisation with Levenberg-Marquardt. + +The functions estimate model parameters by maximising the log-likelihood (or free energy) through iterative optimisation. The algorithms employ the Levenberg-Marquardt +method with dynamic updates of hyperparameters (observation variance and damping factor). + +Usage: + +Given a fully specified DCM, do: + +''' +M = aFitDCM(DCM); +M.aloglik(num_iter) ; <-- log likelihood estimation or +M.aloglikFE(num_iter) ; <-- free enery estimation + +% to re-run / add more iterations: +M.update_parameters(M.Ep) +M.aloglikFE(num_iter) + +% and to access posteriors: +M.Ep +M.CP +M.F +'''