ellisp/forecastxgb-r-package

Issues

• question: is this package abandoned?

  #45 opened 4 years ago by ssh352
  0

• Maxlag is negative for short time series

  #44 opened 6 years ago by Ic3fr0g
  0

• forecast.xgbar() is inaccessible with R 3.5.0

  #43 opened 7 years ago by SpencerDH
  1

• Training period

  #41 opened 7 years ago by MRibeiro2107
  6

• How to predict step by step

  #42 opened 7 years ago by xu-wang11
  1

• MAXLAG XREGS

  #40 opened 7 years ago by NahuelGrasso
  0

• Better way to choose maxlag

  #14 opened 8 years ago by ellisp
  1

• Hyperparameter tuning for xgboost?

  #39 opened 7 years ago by Ic3fr0g
  0

• forecasts with xreg with 1 period forward don't work

  #38 opened 7 years ago by ellisp
  0

• Install error in MacOS

  #37 opened 7 years ago by xizepu
  0

• possibly use designmatrix package

  #8 opened 8 years ago by dashaub
  4

• the meaning of the model reults

  #34 opened 8 years ago by xiaoye10
  2

• lambda problem

  #36 opened 8 years ago by ellisp
  0

• forecast doesn't work with decompose seas_method

  #35 opened 8 years ago by ellisp
  1

• When I run the demo code, an error occurried - "result would be too long a vector"

  #33 opened 8 years ago by chaojieji
  0

• Handle NAs

  #13 opened 8 years ago by ellisp
  0

• Better handling of trends

  #25 opened 8 years ago by ellisp
  1

• odd behaviour when lambda = 1, decompose, differencing

  #30 opened 8 years ago by ellisp
  1

• decompose with non-complete cycles creates NA problems

  #31 opened 8 years ago by ellisp
  0

• "decompose" method doesn't work well in combination with differencing

  #32 opened 8 years ago by ellisp
  0

• daily data - better treatment of series with high value of frequency

  #22 opened 8 years ago by ellisp
  2

• change argument name from maxlag to p, P, d or D or similar?

  #17 opened 8 years ago by ellisp
  0

• thorough test of different ways of choosing lambda

  #27 opened 8 years ago by ellisp
  1

• more ways of dealing with seasonality

  #26 opened 8 years ago by ellisp
  1

• Bug - cannot handle non-integer frequency

  #29 opened 8 years ago by SpencerDH
  2

• thorough test of the different methods of dealing with seasonality

  #28 opened 8 years ago by ellisp
  0

• Error in x[, maxlag + 1] <- time(y2) & Error in x[, maxlag + 2:f] <- seasons

  #20 opened 8 years ago by mhnierhoff
  6

• better treatment of seasons for short series

  #23 opened 8 years ago by ellisp
  1

• Add Box Cox option

  #19 opened 8 years ago by ellisp
  0

• default value for h in forecast.xgbts()

  #24 opened 8 years ago by dashaub
  1

• short monthly series fail

  #21 opened 8 years ago by ellisp
  1

• change name from xgbts to xgbar?

  #16 opened 8 years ago by ellisp
  0

• thorough tests for the xreg functionality

  #11 opened 8 years ago by ellisp
  2

• Passing params to xgboost

  #18 opened 8 years ago by dashaub
  1

• fails when maxlag = 1

  #15 opened 8 years ago by ellisp
  0

• Test against Tcomp and Mcomp

  #6 opened 8 years ago by ellisp
  1

• add vignette

  #10 opened 8 years ago by ellisp
  0

• Examples in helpfiles

  #12 opened 8 years ago by ellisp
  0

• overfitting

  #7 opened 8 years ago by ellisp
  1

• add travis ci

  #9 opened 8 years ago by ellisp
  0

• add support for xreg

  #3 opened 8 years ago by ellisp
  0

• summary method for xgbts objects

  #1 opened 8 years ago by ellisp
  0

• doesn't work with non-seasonal data

  #2 opened 8 years ago by ellisp
  0

• parallelization of cross validation step

  #4 opened 8 years ago by ellisp
  0

• prediction intervals

  #5 opened 8 years ago by ellisp
  0