From 7a5dd0e8551ab09c5ab8b946f5247586256cc91c Mon Sep 17 00:00:00 2001 From: Vikram Oberoi Date: Mon, 22 Apr 2019 11:04:08 -0700 Subject: [PATCH] Ch. 3 correction: MCMC *not* exploring the space well likely exhibits high autocorr. These sentences suggest that a chain exploring the space well will exhibit high autocorrelation. I think the intent is to convey the opposite. --- Chapter3_MCMC/Ch3_IntroMCMC_PyMC2.ipynb | 2 +- Chapter3_MCMC/Ch3_IntroMCMC_PyMC3.ipynb | 2 +- Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/Chapter3_MCMC/Ch3_IntroMCMC_PyMC2.ipynb b/Chapter3_MCMC/Ch3_IntroMCMC_PyMC2.ipynb index a7fdb028..1f09ba5c 100644 --- a/Chapter3_MCMC/Ch3_IntroMCMC_PyMC2.ipynb +++ b/Chapter3_MCMC/Ch3_IntroMCMC_PyMC2.ipynb @@ -1091,7 +1091,7 @@ "\n", "By the nature of the MCMC algorithm, we will always be returned samples that exhibit autocorrelation (this is because of the step `from your current position, move to a position near you`).\n", "\n", - "A chain that is [Isn't meandering exploring?] exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n", + "A chain that is not exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n", "\n", "This does not imply that a converged MCMC has low autocorrelation. Hence low autocorrelation is not necessary for convergence, but it is sufficient. PyMC has a built-in autocorrelation plotting function in the `Matplot` module. " ] diff --git a/Chapter3_MCMC/Ch3_IntroMCMC_PyMC3.ipynb b/Chapter3_MCMC/Ch3_IntroMCMC_PyMC3.ipynb index 37947ae9..6182a6c6 100644 --- a/Chapter3_MCMC/Ch3_IntroMCMC_PyMC3.ipynb +++ b/Chapter3_MCMC/Ch3_IntroMCMC_PyMC3.ipynb @@ -1073,7 +1073,7 @@ "\n", "By the nature of the MCMC algorithm, we will always be returned samples that exhibit autocorrelation (this is because of the step `from your current position, move to a position near you`).\n", "\n", - "A chain that is [Isn't meandering exploring?] exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n", + "A chain that is not exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n", "\n", "This does not imply that a converged MCMC has low autocorrelation. Hence low autocorrelation is not necessary for convergence, but it is sufficient. PyMC3 has a built-in autocorrelation plotting function in the `plots` module. " ] diff --git a/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb b/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb index 05ce6a59..c1257def 100644 --- a/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb +++ b/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb @@ -1909,7 +1909,7 @@ "\n", "By the nature of the MCMC algorithm, we will always be returned samples that exhibit autocorrelation (this is because of the step `from your current position, move to a position near you`).\n", "\n", - "A chain that is [Isn't meandering exploring?] exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n", + "A chain that is not exploring the space well will exhibit very high autocorrelation. Visually, if the trace seems to meander like a river, and not settle down, the chain will have high autocorrelation.\n", "\n", "This does not imply that a converged MCMC has low autocorrelation. Hence low autocorrelation is not necessary for convergence, but it is sufficient. TFP has a built-in autocorrelation tools as well. " ]