Merge changes for v0.3.0 #86
Merged
Conversation
This file contains 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
Addresses deprecation warnings and updates examples in documentation.
Require julia v0.6
Compat is no longer needed since only julia 0.6 is currently supported.
Drop `Compat` dependency
Fixes #57 (part 1) by joining reports.
Fix joining of codecov flags
Add convex decreasing density estimator for π0
Adds methods and interfaces for the estimation of higher criticism scores and thresholds
Remove leftovers from julia v0.5
- Update details of software packages - Merge papers and software packages
Update and combine references
Simple test case: `@benchmark sortperm(pv) setup=(pv=PValues(rand(1000)))` Before: 3.2ms After: 110μs Reference: 54μs: `@benchmark sortperm(pv) setup=(pv=rand(1000))`
Link comic logo
Update documentation notebooks
Update tests and references for p-value combinations
`digamma` has been deprecated in base.
Fix slow `PValues` operations by using concrete types
Address some deprecations in future julia versions
Drop qvalues in favour of BH-adjusted p-values
Simplify the method structure
* Use faster computation for sum of squares * Use partial sorting for a faster Wilkinson combination * Simplify boundary checks for p-value combinations
* Define return types for p-value combinations: `combine(PValues{T}, ...)` now always returns a p-value of type `T`.
* Add tests for type stability and defined return types for combinations.
- The modification of the p-values is now part of the main `adjust` functions, no additional `_step` functions are needed: This gives us fewer functions to worry about, cleaner code and better performance (e.g. we only have to compute the harmonic number once per adjustment and don't need to recompute the same value for each call of the step function). - `stepup!` and `stepdown!` are now only performing the monotisation with `cummin` and `cummax`; the modification of the p-values happens beforehand. - The new variable names reflect their usage more clearly and consistently.
Fixes `ForwardStop` and `Hommel` adjustment with unsorted p-values. Note that the sorting order of the output is not well-defined for the `ForwardStop` adjustment in case the input has ties (see the test cases).
Test also with julia nightly
Reference citation for the convex decreasing estimator
- Don't use the vectorised min/max for a scalar - Use `clamp` instead of max(min(x)) which is cleaner and faster
Simplify π0 estimators
Update reference for Tippett's combination
Drops the generic `combine` methods with weights: Only the Stouffer method currently supports weights. Unifies the `combine` methods with `StatsBase.Weights` and `Vector` weights to support only an `AbstractVector` input: This handles both existing cases with one method, and provides a more flexible interface for additional types of weights.
Unify methods for weighted Stouffer combination
Use `clamp` for better bounding of estimates
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Version 0.3.0
New Features
HigherCriticismScores) and threshold (HigherCriticismThreshold) estimation (Higher criticism scores and critical value estimation #55)Changes
User-facing changes
Internal changes
clampfor better bounding of estimates (Useclampfor better bounding of estimates #85, Simplify π0 estimators #82)Distributionsinterface (Require julia v0.6 #52)Compatdependency (DropCompatdependency #53)Removed
qvaluesin favour of BH-adjusted p-values (Drop qvalues in favour of BH-adjusted p-values #73).Support