mlr3mbo
mlr3mbo picked its defaults from a benchmark study, not from taste
Comparison · Analytics
discrim vs embed
A side-by-side editorial comparison of discrim and embed — release velocity, themes, recent moves, and the top alternatives to consider.
Shared themes:tidymodels
discrim vs embed: at a glance
What is discrim?
discrim settled into a thin engine shim after handing its model definitions to parsnip.
discrim is the parsnip extension for discriminant analysis, exposing linear, quadratic, flexible and regularized variants through the tidymodels interface. Its model definition functions moved into parsnip itself in 0.2.0, leaving this package as the engine and prediction layer. The two most recent releases are a documentation fix and a single prediction bug.
What is embed?
embed keeps adding encoding steps while shedding its deep-learning dependencies
embed supplies recipes steps that turn categorical predictors into numeric representations — likelihood encoding, UMAP projection, string-distance collapsing. The 1.1.x line made UMAP arguments tunable and moved keras and tensorflow out of hard dependencies; 1.2.0 added analytical likelihood encoding with partial pooling and retired step_feature_hash() in favor of textrecipes.
discrim vs embed: editorial side-by-side
discrim
0.0ANALYTICS
discrim settled into a thin engine shim after handing its model definitions to parsnip.
◆ Current state
discrim is the parsnip extension for discriminant analysis, exposing linear, quadratic, flexible and regularized variants through the tidymodels interface. Its model definition functions moved into parsnip itself in 0.2.0, leaving this package as the engine and prediction layer. The two most recent releases are a documentation fix and a single prediction bug.
◆ Where it's heading
The package built out method coverage early, adding quadratic discriminant analysis in 0.1.2, the sda and sparsediscrim engines in 0.1.3 and case weights in 1.0.0, then stopped growing. Handing definitions upstream to parsnip in 0.2.0 confirmed the shape: discrim is where engines are wired, not where the API lives. Cadence since 2022 is roughly one small fix a year.
◆ Prediction
Nothing in the entries points to new methods or engines; the next release is most likely another CRAN or prediction-path fix.
embed
0.0ANALYTICS
embed keeps adding encoding steps while shedding its deep-learning dependencies
◆ Current state
embed supplies recipes steps that turn categorical predictors into numeric representations — likelihood encoding, UMAP projection, string-distance collapsing. The 1.1.x line made UMAP arguments tunable and moved keras and tensorflow out of hard dependencies; 1.2.0 added analytical likelihood encoding with partial pooling and retired step_feature_hash() in favor of textrecipes.
◆ Where it's heading
Two quiet directions run through these releases. One is making the steps tunable rather than fixed, so they participate properly in tidymodels grids. The other is boundary maintenance: heavy dependencies pushed to Suggests, overlapping steps handed to the package that owns them. Recent releases are thin and fix-driven.
◆ Prediction
Expect further consolidation with textrecipes over which package owns which encoding step, and continued upkeep against xgboost and uwot releases rather than new step families.
Alternatives to discrim and embed
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 discrim or embed.
loo
loo keeps rewriting the diagnostics Bayesian modellers read off model comparison
mlr3fselect
mlr3fselect turned feature selection into an asynchronous, distributable job
lime
lime survives on compatibility patches years after its research moment
mlr3measures
mlr3measures is systematically retrofitting sample weights across every metric
mlr3cluster
mlr3cluster went from a handful of clusterers to covering the field
See all discrim alternatives → · See all embed alternatives →
Recent activity from discrim and embed
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
- 6mo agoembedstep_umap() zero-component bug fixed
- 8mo agoembedCompatibility with all xgboost versions
- 8mo agodiscrimFix for FDA models failing at prediction time
- 11mo agoembedstep_lencode() adds analytical likelihood encoding with pooling
- 11mo agodiscrimDocumentation links updated to stay on CRAN
- 1y agoembedUMAP initial and target_weight become tunable
- 2y agoembedkeras and tensorflow moved to Suggests
- 2y agoembedstep_collapse_stringdist() returns factors
- 4y agodiscrimCase weights enabled for flexible and linear discriminant models
- 4y agodiscrimModel definitions moved upstream into parsnip
- 5y agodiscrimsda and sparsediscrim engines added for LDA and QDA
- 5y agodiscrimdiscrim_quad() added; package relicensed to MIT ⚡
Frequently asked questions
What is the difference between discrim and embed?
Both compete on the same themes — tidymodels — within Analytics. discrim and embed 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 discrim better than embed?
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. discrim and embed 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 discrim?
Top discrim alternatives in Analytics are ranked by recent ship velocity. Browse the "discrim alternatives" section above for the current picks, or visit /alternatives/discrim for the full list with editorial commentary on each.
What are the best alternatives to embed?
Top embed alternatives in Analytics are ranked by recent ship velocity. Browse the "embed alternatives" section above for the current picks, or visit /alternatives/embed for the full list with editorial commentary on each.