Code

Selective drop#1

Code

Merged

ccomkhj merged 28 commits intomasterccomkhj/constrainedML:masterfrom selective_dropccomkhj/constrainedML:selective_dropCopy head branch name to clipboardNov 13, 2023

+3,297-389

Commits

Commits on Nov 3, 2023

• restructure; __init__ to each module

  ccomkhj

  committed
  ac65d5e

  Copy full SHA for ac65d5e

• introduce unit test for MultiConstraintLinearRegression

  ccomkhj

  committed
  eb41bd6

  Copy full SHA for eb41bd6

• introduce base class

  ccomkhj

  committed
  6ffc9f2

  Copy full SHA for 6ffc9f2

• inherit from base

  ccomkhj

  committed
  87bdd82

  Copy full SHA for 87bdd82

• modulizing shape checks

  ccomkhj

  committed
  9fb6c79

  Copy full SHA for 9fb6c79

• docstring; formatting; private method;

  ccomkhj

  committed
  f90dd46

  Copy full SHA for f90dd46

• delete overlap methods from super class;

  ccomkhj

  committed
  ba8de29

  Copy full SHA for ba8de29

• introduce test; comparision between single and multi constrained LR; shape check;

  ccomkhj

  committed
  7f03cf9

  Copy full SHA for 7f03cf9

• update tutorial of multiconstrainedLR based on updates

  ccomkhj

  committed
  044ed75

  Copy full SHA for 044ed75

Commits on Nov 6, 2023

• build constrained_mlp; check if it works samely with sklearns mlp

  ccomkhj

  committed
  6425224

  Copy full SHA for 6425224

• introduce clipping coefs

  ccomkhj

  committed
  f74bd21

  Copy full SHA for f74bd21

• introduce multi coefficients

  ccomkhj

  committed
  0bfb432

  Copy full SHA for 0bfb432

• introduce test for mlp

  ccomkhj

  committed
  257f725

  Copy full SHA for 257f725

• fix typo and error

  ccomkhj

  committed
  c14c9d9

  Copy full SHA for c14c9d9

Commits on Nov 7, 2023

• formatting

  ccomkhj

  committed
  92c5ad1

  Copy full SHA for 92c5ad1

• gradient update only once per iteration

  ccomkhj

  committed
  e5b05bc

  Copy full SHA for e5b05bc

• include loss to track model learning

  ccomkhj

  committed
  d446157

  Copy full SHA for d446157

Commits on Nov 8, 2023

• grad update after beta iteration; equal to basemodel

  ccomkhj

  committed
  50b92f8

  Copy full SHA for 50b92f8

• lbfgs works with clipping

  ccomkhj

  committed
  384cf5f

  Copy full SHA for 384cf5f

• include training loss in mlp

  ccomkhj

  committed
  978d1a9

  Copy full SHA for 978d1a9

• update docs

  ccomkhj

  committed
  84a1e04

  Copy full SHA for 84a1e04

• minor font

  ccomkhj

  committed
  ecb6544

  Copy full SHA for ecb6544

• minor font2

  ccomkhj

  committed
  4cb5b8b

  Copy full SHA for 4cb5b8b

Commits on Nov 10, 2023

• introduce selective drop LR

  ccomkhj

  committed
  d796262

  Copy full SHA for d796262

• introduce test for selecctive drop LR

  ccomkhj

  committed
  c97630f

  Copy full SHA for c97630f

• mlp with bounds to directly optimizer

  ccomkhj

  committed
  65d29bd

  Copy full SHA for 65d29bd

• introduce selective drop module

  ccomkhj

  committed
  575100b

  Copy full SHA for 575100b

Commits on Nov 13, 2023

• introduce selective_drop_positive_lr; utilize positive flag from the native lr in sklearn

  ccomkhj

  committed
  95b2791

  Copy full SHA for 95b2791