Add stochastic volatility model from Hoffman & Gelman 2014#326
Open
colinpochart2 wants to merge 16 commits into
Open
Add stochastic volatility model from Hoffman & Gelman 2014#326colinpochart2 wants to merge 16 commits into
colinpochart2 wants to merge 16 commits into
Conversation
Collaborator
|
Excellent! @JTorgander could you do a review? |
JTorgander
requested changes
Jun 4, 2026
| } | ||
|
|
||
| // Term from marginalizing out tau analytically: | ||
| // target += -((T + 1) / 2.0) * log(0.01 + 0.5 * sum((log s_i - log s_{i-1})^2)) |
Collaborator
There was a problem hiding this comment.
This line can be removed
Comment on lines
+1
to
+4
| functions { | ||
|
|
||
| } | ||
|
|
Collaborator
There was a problem hiding this comment.
This block can be removed
| target += -0.01 * s[1]; | ||
|
|
||
| // Likelihood: (log y_i - log y_{i-1}) / s_i ~ t_nu | ||
| // Add -log(s_i) for each term because of the Jacobian of standardization. |
Collaborator
There was a problem hiding this comment.
Where is this happening in the code?
| } | ||
|
|
||
| transformed data { | ||
| vector[T-1] log_returns; // log y_i - log y_{i-1} for i = 2, ..., T |
Collaborator
There was a problem hiding this comment.
This comment can be removed.
Collaborator
There was a problem hiding this comment.
I think we need to include information about the data source, time range etc
| "name": "stochastic_volatility", | ||
| "keywords": ["time_series", "stochastic_volatility", "marginalized"], | ||
| "title": "Stochastic Volatility Model with Marginalized Precision", | ||
| "description": "Stochastic volatility model from Hoffman & Gelman (2014, NUTS paper, p. 1614-1615). Daily log-returns follow a Student-t distribution with time-varying scale s_i. The log-scale follows a Gaussian random walk with precision tau, which is marginalized analytically. Resulting posterior is (T+1)-dimensional over (s, nu).", |
Collaborator
There was a problem hiding this comment.
Suggested change
| "description": "Stochastic volatility model from Hoffman & Gelman (2014, NUTS paper, p. 1614-1615). Daily log-returns follow a Student-t distribution with time-varying scale s_i. The log-scale follows a Gaussian random walk with precision tau, which is marginalized analytically. Resulting posterior is (T+1)-dimensional over (s, nu).", | |
| "description": "Stochastic volatility model from Hoffman & Gelman (2014, NUTS paper, p. 1614-1615). Daily log-returns follow a Student-t distribution with time-varying scale s_i. The log-scale follows a Gaussian random walk with precision tau, which is marginalized analytically. Resulting posterior is T-dimensional over (s, nu), where T is the number of data points.", |
Collaborator
There was a problem hiding this comment.
It is a bit confusing to the user that the raw data has 3021 data points whereas the json data file 3000.
Collaborator
There was a problem hiding this comment.
I think it makes more sense to store this data under the name sp500_[time-range]_T and then name the posterior sp500_[time-range]_T-stochastic_...
Co-authored-by: JTorgander <55882444+JTorgander@users.noreply.github.com>
- info.json: posterior is T-dimensional over (s, nu), not (T+1) - stochastic_volatility.stan: remove empty functions block and stale comments
colinpochart2
force-pushed
the
add-stochastic-volatility
branch
from
fba05bd to
bac96c0
Compare
Co-authored-by: JTorgander <55882444+JTorgander@users.noreply.github.com>
Co-authored-by: JTorgander <55882444+JTorgander@users.noreply.github.com>
…_volatility.json Co-authored-by: JTorgander <55882444+JTorgander@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds the stochastic volatility model from Hoffman & Gelman (2014, NUTS paper).
PS: reference posterior draws are not yet included - the HMC run is computationally demanding. Happy to follow
maintainer guidance on the sampling configuration.