This announces the v0.1.4 release of NearestNeighborModels.jl.
NearestNeighborModels.jl builds on @kristoffer.carlsson NearestNeighbors package and earlier contributions from @tlienart residing at MLJModels.jl. The package implements K-nearest neighbor regressors(KNNRegressor, MultitargetKNNRegressor) and classifiers(KNNClassifier, MultitargetKNNClassifier), for use with the MLJ machine learning framework. In addition to this a variety of weighing kernels are provided including those from the paper Geler et.al (2016):
Comparison of different weighting schemes for the kNN classifier on time-series data.
9 Likes