Changelog
This changelog includes end-user changes only. For internal development notes, see the repository changelog.
[5.1.0]
Changed
-
Breaking change:
OnlineLowess.addPointreturnsAddPointResultwithOnlinePointStatus(FILLING/READY/DROPPED) instead of a bareOptional<PointResult>. -
Online fits sort the sliding window by
xbefore each update so non-monotonic streams stay correct. -
StreamingLowess.processChunkrejects chunks that are not non-decreasing inxwhen merged with the previous overlap; feed ascending chunks or pre-sort the series.
Fixed
-
Distinguishing Online
addPointfilling vsmissing="drop"viaOnlinePointStatus/AddPointResult. -
Streaming no longer panics when derivatives were requested but missing; it throws instead.
-
Extend/Zero boundary pads use unit spacing when leading/trailing
xties would otherwise stack pads. -
Closing retained
PredictModelhandles throughResult(AutoCloseable) with take-once ownership and aCleanersafety net on native-owning types. -
Validating equal
x/ylengths andcustomWeightsagainstybefore JNI.
[5.0.0]
Added
-
Enabled the optional wgpu DirectX 12 backend for GPU-enabled Windows builds. DXC is loaded dynamically with an FXC fallback, avoiding eager imports of
dxcompiler.dllanddxil.dll. -
Added GLES for Android GPU builds with a target-scoped
wgpufeature; Windows continues to omit the GLES-only loader imports. -
Added grouped
outputs(…),cv(CVOptions…), and prediction-output APIs while preserving native option mapping. -
Added
IntervalsOptionsfor grouped confidence/prediction levels and residual-bootstrap intervals inLowess,StreamingLowess, full-updateOnlineLowess, andPredictModel.predict. -
Added an "Alternative Software" guide comparing
fastlowesswith Apache Commons Math’sLoessInterpolator.
Changed
-
Clarified that Batch
residualSd()is1.4826 * MAD, while Streaming reports the cumulative sample standard deviation of emitted residuals. -
Breaking change: replaced individual
return…output setters and prediction flags withoutputs(…)for Batch, Streaming, Online, and prediction options. -
Breaking change: replaced flat interval setters and
PredictOptionsinterval levels withintervals(IntervalsOptions), and replaced flat CV setters andCVOptions.seedwithcv(CVOptions)plus one outerseed(long)shared by CV and bootstrap.PredictOptionshas its ownseedfor prediction-time bootstrap.
Fixed
-
Fixed CPU Gaussian standard errors for fitted and queried values to use the full unbounded kernel support.
-
Made even-sample medians and mean-absolute residual scaling overflow-resistant for large finite values, and centered Batch/Streaming R-squared accumulation to preserve one-ULP response variation at large offsets.
-
Made Batch RMSE, MAE, and R-squared reductions scale-safe for large finite values, preserved bisquare downweighting when tuned scales exceed the numeric range, and normalized custom weights before summation in local and all-tied fits.
-
Kept AIC finite when raw residual-square sums overflow, and made local/all-tied WLS invariant to common scaling of large finite case weights.
-
Fixed global OLS fits treating predictor values with a large offset as degenerate; translated inputs now retain their fitted slope.
-
Fixed fraction-1 global fits ignoring custom weights, including when Batch sorts observations by x.
-
Fixed Batch
missing = "drop"accepting custom weights with a length different from the original input; weights are validated before rows are dropped. -
Fixed standard errors for global weighted fits to account for observation weights and weighted prediction leverage.
-
Fixed local Batch and retained-model prediction standard errors ignoring
custom_weights; local SE moments now include the per-observation case weights. -
Fixed global fits with all-zero custom weights to honor the configured zero-weight fallback policy.
-
Fixed fraction-1 global fits ignoring configured robustness iterations; they now reweight observations and report iterations used.
-
Fixed Batch cross-validation candidate fits ignoring
custom_weights; K-fold CV now rejects more folds than observations instead of returning zero scores. -
Fixed Streaming and Online accepting invalid
auto_convergetolerances; Online also rejects invalid explicitdeltavalues while retaining NaN as its default sentinel. -
Fixed GPU Batch fits misaligning fitted values with unsorted inputs; results now preserve input and requested sorted order.
-
Fixed GPU Batch silently ignoring
custom_weights; GPU fit and CV candidate kernels now apply them directly. -
Reduced GPU adapter buffer requirements from 30 storage/32 total buffer bindings to 7 storage/8 total per shader stage by using per-compute-pipeline resource layouts.
-
Fixed Online incremental mode accepting positive
deltaandauto_convergesettings it cannot use; unsupported combinations now error. -
Fixed grouped intervals discarding distinct confidence and prediction levels; each requested coverage is now applied independently.
-
Fixed retained-model handles leaking when JNI result conversion fails and array-read failures being hidden as empty inputs.
-
Fixed invalid k-fold counts and negative Streaming/Online iterations being silently coerced; CV grids and explicit custom-weight arrays must now be non-empty/valid.
-
Fixed GPU installation selecting unsupported libc/architecture targets or accepting incompatible local libraries; supported targets and a matching GPU ABI marker are now required.
-
Serialized operations on each Java model instance to protect mutable native state during concurrent calls.
-
Fixed the default
boundaryPolicy("extend") letting synthetic boundary points bias the shared robustness scale estimate used to reweight every point, compounding across robustness iterations. -
Improved agreement with R/Cleveland on sparse, asymmetric, and high-iteration fits by aligning robustness stopping, local-linear degeneracy handling, neighborhood traversal, delta interpolation, and weighted accumulation.
-
Fixed zero-radius neighborhoods dropping tied observations or accumulating normalized weights in a different order under robust fits.
-
Fixed Gaussian fits clipping the unbounded kernel to the neighbor window and flooring far-tail weights; all observations now contribute under the standard Gaussian formula.
4.1.0
Added
-
Added bundled native libraries for the Java binding across 8 platforms, with runtime musl detection and auto-extraction.
-
Added
retainModelandResult.predictModel()for prediction. -
Added
returnDerivative(boolean)toOptionsandOnlineOptions. -
Added standard errors and confidence/prediction intervals to
StreamingOptionsandOnlineOptions; online intervals requireupdateMode("full").
4.0.0
Added
-
Added
returnSortedtoOptions. -
Added a
missingoption toOptions.Builder, shared byOptions,StreamingOptions, andOnlineOptions, to control non-finite input handling. -
Added
FastLowess.installGpu()to download a prebuilt GPU library; the Java runtime can load it from a directory set with-Dfastlowess.native.dir. -
GPU installers now accept a path to a local GPU artifact, and prebuilt GPU artifacts are available for Linux and Windows ARM64.
Changed
-
Breaking change: removed
returnDiagnostics,returnResiduals, andparallelfromOnlineOptions.Builder; these options were accepted but had no effect in online mode. -
Breaking change: removed
confidenceIntervalsandpredictionIntervalsfromOnlineOptions.BuilderandStreamingOptions.Builder. -
Breaking change:
StreamingOptions.Builder.overlap()now defaults dynamically tochunk_size / 10, clamped to[1, chunk_size - 10], instead of a fixed500. -
Improved the Java API documentation, including the overlap default.