Release 1.4-beta.16

Additions

1. KFilter and mcmc objects: Preliminary implementations of Kalman filtering thanks to @amitkumarj441
2. Added overloaded apply method to CovarianceFunction interface allowing user to create arbitrary tunable kernels (using GlobalOptimizer implementations like GridSearch and CoupledSimulatedAnnealing).

Work in Progress

1. Improvement of Kalman filtering implementation making it more modular and in sync with DynaML Model[PQR] API cc: @amitkumarj441 .

Release 1.4-beta.15

Additions

1. Composition operations defined for ReversibleScaler class
2. Added AutoEncoder as a subclass of ReversibleScaler

Errata

1. Corrected typo in name of BackPropagation class

Release 1.4-beta.14

Improvements

1. Added ++ append operator to RegressionMetrics, BinaryClassificationMetrics and MultiRegressionMetrics
2. Added implementation for generatePlots function in MultiRegressionMetrics

Release 1.4-beta.13

Additions

1. Added gaussianScaling and minMaxScaling to DynaMLPipe

Improvements

1. Changed return types of gaussianScalingTrainTest and minMaxScalingTrainTest in DynaMLPipe.

Release 1.4-beta.12

Additions

1. Added new implementations of [0,1] and Gaussian scalers, which return the appropriate ReversibleScaler instances.

Release 1.4-beta.11

Improvements

1. Added MultiRegressionMetrics class for calculating regression performance on multi-output tasks.

Release 1.4-beta.10

Additions

1. Scaler API, ability to write gaussian, min-max scaling functions and filters
2. Haar Wavelet transform, along with inverse wavelet transform

Modifications

1. Moved around graph package classes

Release 1.4-beta.9

Improvements

1. Added Haar wavelet transform
2. Aggregating build

Release 1.4-beta.8

Changes

Converted to multi module sbt distribution.

Release 1.4-beta.7

Bug fix to Polynomial kernel.