rjd3highfreq
rjd3highfreq ships whatever the Java side ships, and only occasionally says what that was.
Comparison · Analytics
broadcast vs kernelshap
A side-by-side editorial comparison of broadcast and kernelshap — release velocity, themes, recent moves, and the top alternatives to consider.
broadcast vs kernelshap: at a glance
What is broadcast?
broadcast is filling in NumPy-style array broadcasting for R, operator by operator.
broadcast brings dimension-broadcasting semantics to R arrays and lists — elementwise operations between arrays of mismatched shape, plus casting methods between hierarchical lists and dimensional structures. It reached CRAN in September 2025 and has released roughly monthly since, accumulating operators (nor, nand, longest common substring), casting methods (cast_shallow2atomic, cast_hier2dim, hiernames2dimnames), and helpers (vector2array, undim, mbroadcasters).
What is kernelshap?
kernelshap makes permutation SHAP practical past eight features, then fixes the kernel weights it had wrong.
kernelshap computes model-agnostic SHAP values in R through Kernel SHAP, permutation SHAP and an exact additive explainer. Version 0.8.0 added a sampling permutation-SHAP algorithm with standard errors and early stopping, lifting the practical feature ceiling past what the exact method allows. Version 0.9.0 then corrected a bug in how kernel weights were computed — exact Kernel SHAP now agrees with exact permutation SHAP — and moved parallelism from foreach to doFuture.
broadcast vs kernelshap: editorial side-by-side
broadcast
0.0ANALYTICS
broadcast is filling in NumPy-style array broadcasting for R, operator by operator.
◆ Current state
broadcast brings dimension-broadcasting semantics to R arrays and lists — elementwise operations between arrays of mismatched shape, plus casting methods between hierarchical lists and dimensional structures. It reached CRAN in September 2025 and has released roughly monthly since, accumulating operators (nor, nand, longest common substring), casting methods (cast_shallow2atomic, cast_hier2dim, hiernames2dimnames), and helpers (vector2array, undim, mbroadcasters).
◆ Where it's heading
The package is in its post-launch consolidation year, and the release notes read accordingly: roughly half of each entry is a consistency correction rather than an addition. Zero-length results now carry the right type, comparison operators accept integer and logical inputs, the comment attribute survives operations, and the nand operator was found to be wrongly defined against C++ short-circuit evaluation. That ratio is what a young package looks like while its edge cases are being found.
◆ Prediction
Expect more operators and casting methods on the same cadence, with continued type-consistency corrections as users exercise unusual input combinations. Nothing in the entries points at an architectural change.
kernelshap
0.0ANALYTICS
kernelshap makes permutation SHAP practical past eight features, then fixes the kernel weights it had wrong.
◆ Current state
kernelshap computes model-agnostic SHAP values in R through Kernel SHAP, permutation SHAP and an exact additive explainer. Version 0.8.0 added a sampling permutation-SHAP algorithm with standard errors and early stopping, lifting the practical feature ceiling past what the exact method allows. Version 0.9.0 then corrected a bug in how kernel weights were computed — exact Kernel SHAP now agrees with exact permutation SHAP — and moved parallelism from foreach to doFuture.
◆ Where it's heading
Two concerns drive this package: making exact methods reach further, and being demonstrably right. The first shows in the additive explainer, the optional background dataset and the sampling permutation algorithm; the second in unit tests written against Python's shap, credited fixes from outside contributors, and a willingness to ship a correctness fix that changes numbers people have already published. Speed work runs continuously underneath — direct solves replacing the Moore-Penrose pseudo-inverse, roughly 10% less memory.
◆ Prediction
The 0.6.0 and 0.7.0 notes each promised a stable 1.0.0 that has not arrived; with the weighting bug fixed and parallelism reworked, a 1.0 release is the most plausible next step.
Alternatives to broadcast and kernelshap
Other Analytics products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either broadcast or kernelshap.
audubon
audubon's release feed is almost entirely Renovate bumping the JavaScript toolchain behind its Japanese text splitter.
affiner
affiner is quietly turning a grid transformation helper into a small computational geometry library.
ageproR
ageproR spent two years chasing a moving file format, then added the recruitment models that justify the effort.
r-ledger
ledger adds a Rust toolchain fallback, so beancount imports work whether or not the Python tooling is installed.
gridpattern
gridpattern keeps widening its catalogue, and the newest patterns finally use the device's own line rendering.
See all broadcast alternatives → · See all kernelshap alternatives →
Recent activity from broadcast and kernelshap
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
- 2mo agobroadcastnor and longest-common-substring operators added; nand corrected
- 5mo agobroadcastcheckNULL, checkNA and ecumprob added
- 8mo agobroadcastZero-length results and attribute preservation made consistent
- 9mo agobroadcastacast dimnames bug fixed; casting and helper surface widens
- 10mo agobroadcastrecurse_classed replaced by recurse_all in casting methods
- 11mo agobroadcastTitle case fixed for CRAN submission
- 1y agokernelshapKernel weight bug fixed; parallelism moves to doFuture
- 1y agokernelshapSampling permutation SHAP with standard errors ⚡
- 1y agokernelshapBackground data now optional; ranger survival support
- 2y agokernelshapFactor-valued predictions dropped
- 2y agokernelshapadditive_shap() explains additive models exactly
- 2y agokernelshapFaster on plain data.frames
Frequently asked questions
What is the difference between broadcast and kernelshap?
They serve adjacent needs but don't currently overlap on shipped themes. broadcast and kernelshap are shipping at a similar cadence (velocity 0.0 vs 0.0, both within Sparkpulse's "active" band). See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Is broadcast better than kernelshap?
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. broadcast and kernelshap are shipping at a similar cadence (velocity 0.0 vs 0.0, both within Sparkpulse's "active" band). For your specific use case, the alternatives sections above list other Analytics products to evaluate alongside.
What are the best alternatives to broadcast?
Top broadcast alternatives in Analytics are ranked by recent ship velocity. Browse the "broadcast alternatives" section above for the current picks, or visit /alternatives/broadcast-r for the full list with editorial commentary on each.
What are the best alternatives to kernelshap?
Top kernelshap alternatives in Analytics are ranked by recent ship velocity. Browse the "kernelshap alternatives" section above for the current picks, or visit /alternatives/kernelshap for the full list with editorial commentary on each.