Issues filed for topepo/caret

View Full Project
Do you use caret? Leave a review!

Rate of open issues in the last 60 days

caret open issues (View Closed Issues)
  • about 2 years Error in train function of caret "Error in `[.data.frame`(data, , all.vars(Terms), drop = FALSE) : undefined columns selected"
  • about 2 years Wrong reference in http://topepo.github.io/caret/model-training-and-tuning.html
  • about 2 years Feature request: resamples-type function for predictions on test set
  • about 2 years xgboost class probabilities inaccurate if `objective = "binary:logitraw"`
  • about 2 years Breaking Backward Compatibility
  • about 2 years xgbtree for poisson regression with offset
  • about 2 years Gini and logLoss Goodness of Fit function
  • about 2 years [New Model]: kernDeepStackNet: Kernel Deep Stacking Networks
  • about 2 years [New Model]: monmlp: Monotone Multi-Layer Perceptron Neural Network
  • about 2 years Ability to send additional data to trainControl's summaryFunction
  • over 2 years Sparse Matrix in caret
  • over 2 years feature request: fast frugal trees
  • over 2 years createFolds is very slow when y is a character with many values
  • over 2 years Feature proposal - multiple input mulitple output
  • over 2 years Feature Request - Label K-fold cross validation
  • over 2 years Feature suggestion - LIME
  • over 2 years scaling preds to original outcome distribution when down/upsampling
  • over 2 years parallelizing preprocess
  • over 2 years Feature Request: place ... in evalSummaryFunction() for train()
  • over 2 years Feature request: rBayesian
  • over 2 years Feature proposal: time partition for time series classification/regression
  • over 2 years Change default arguments of ROC computation (pROC) in the RFE process (caret)
  • over 2 years Y-aware PCA for preprocessing
  • over 2 years [New Model]: kinn (an Implementation of 'kinn' Algorithm, a Graph Based Regression Model)
  • over 2 years [New Model]: fuzzyforest
  • over 2 years Add warning with OOB and custom summary metric
  • over 2 years preProcess zv and nzv for factors
  • over 2 years LLVM static code analysis shows ten bugs
  • over 2 years look into deep learning package
  • over 2 years write makefile for html builds
  • over 2 years Intercept xgbTree parameters
  • over 2 years ROC based variable importance error
  • almost 3 years custom holdout indexes: train reports wrong samples, calculations are fine
  • almost 3 years Error with nb from klaR package
  • almost 3 years Apply YJ transform during adaptive resampling
  • about 3 years evaluate addition of mxnetr in model list
  • about 3 years [new models] glinternet package -- learning interactions using Group-Lasso
  • about 3 years Sparse Linear Regression using Nonsmooth Loss Functions and L1 Regularization (flare)
  • about 3 years High-Dimensional Regression and CAR Score Variable Selection
  • about 3 years GBM fails when doing quantile regression
  • about 3 years PCA/spatial sign inconsistencies
  • about 3 years xgboost vs xgbtree
  • about 3 years Investigate mxnet for deep learning
  • about 3 years createDataPartition creates approximate splits
  • about 3 years Add H20 models
  • about 3 years python knn fails with formula method
  • about 3 years mixed models toys for lme4 model inclusion
  • about 3 years bucketing with preProcess
  • about 3 years parallel seeds
  • about 3 years have na.pass insert missing values in predict.train
  • over 3 years Add ability to optimize probability thresholds for class imbalances without having to create a custom model
  • over 3 years How to create data partitions using multiple categorical and numerical columns?
  • over 3 years createFolds with strata
  • over 3 years Getting Confidence & Prediction Intervals from predict.train
  • over 3 years Use data.tables internally to store data?
  • over 3 years investigate the pCovR package
  • over 3 years have the bag function use other sampling methods
  • over 3 years add a metric option to the resamples function
  • over 3 years investigate sgd
  • over 3 years Failures with dev version of testthat
  • over 3 years investigate trimming models and pre-processing objects during resampling
  • over 3 years investigate sgPLS: Sparse Group Partial Least Square Methods
  • over 3 years investigate bootsPLS
  • over 3 years consider adding PRIM
  • over 3 years GAM models in caret do not support Poisson family models
  • almost 4 years more flexibility in pre-processing method call
  • almost 4 years add notes to custom model HTML page
  • almost 4 years look at anfis
  • almost 4 years Investigate irace package as an alternative to tune grids
  • almost 4 years add type = "none" resampling to gafs and safs
  • almost 4 years Add Regression Tests to travis
  • about 4 years Add Nested Cross-Validation
  • about 4 years Sparse Multi-category Angle-Based Large-Margin Classifiers
  • about 4 years error in predict.train
  • about 4 years fix median impute in preProcess
  • about 4 years Add inverse hyperbolic sine transform to preProcess
  • about 4 years write some basic unit tests
  • about 4 years improve sbf score function documentation
  • about 4 years add ordinalgmifs
  • about 4 years parallel option for filterVarImp
  • over 4 years rfe saves predictions incorrectly
  • over 4 years Adapt train for survival models
  • over 4 years investigate FADA package for pre-processing
  • over 4 years make class for holdout predictions
  • over 4 years refine svm rbf grid
  • over 4 years consider moving `metric` in rfe() and sbf() to the control function
  • over 4 years in predict.train, subset on features
  • over 4 years add bootstrap 632+ estimator
  • over 4 years add an `extract` class
  • over 4 years update documentation about `returnResamp`
  • over 4 years Issue a warning with `update.train`
  • over 4 years Investigate allowing prediction function to return multiple values
  • over 4 years Investigate deep learning packages

caret closed issues

  • about 2 years (re)add documentation for RMSE and R2
  • about 2 years BRT train error
  • about 2 years corr in preProcess
  • about 2 years confusionMatrix print formatting broken in version 6.0-72
  • about 2 years Unnamed vector returned by mnLogLoss breaks model selection in training
  • about 2 years Add RANDOM ROTATION ENSEMBLE TECHNIQUE to caret
  • about 2 years Bug: train with rlm only works with factors or characters if all levels would be legal as column names
  • about 2 years classProbs=TRUE causes error in earth model
  • about 2 years Feature Selection and Parameter Tuning
  • about 2 years Something is wrong; all the RMSE metric values are missing
  • about 2 years Override the default behavior of "pca" in preProcess
  • about 2 years Fuzzy methods throw warnings about data range
  • about 2 years Error occured when using prSummary for imbalanced data
  • about 2 years non-default parameters in nzv method (preProcess function)
  • about 2 years varImp includes regressors that are explicitly excluded from model
  • about 2 years xgboost with caret for multiclass classification-- The following parameters were provided multiple times: objective
  • about 2 years 'skip' in createTimeSlices()
  • about 2 years Reproducibility of createFolds()
  • about 2 years Cross-validated rfe: how is the final "best" feature set selected?
  • about 2 years [Documentation Enhancement]: Include reference in "Random Hyperparameter Search"...
  • about 2 years R caret: rfe nnet โ€œundefined columns selectedโ€
  • about 2 years sbf() returns anomalous results with factors
  • about 2 years Number of data points used to estimate lambda in Box-Cox transformations
  • about 2 years train function does not work the same in current version
  • about 2 years Error in na.fail.default and #461 fix
  • about 2 years Model request glm.nb
  • over 2 years What is the varImp mathematical calculation expression?
  • over 2 years memory size problem in caret / caretEnsemble
  • over 2 years New feature: bumping
  • over 2 years Problem with predict.train after R and caret package upgrade.
  • over 2 years RWeka classifier hyperparameters
  • over 2 years xgbTree tuning options
  • over 2 years Confusion Matrix bug?
  • over 2 years na.action default ignored in train
  • over 2 years Inappropriate network architecture of one neuron per layer in dnn
  • over 2 years avMxnet is not a valid caret model
  • over 2 years Error while trying to print gam model.
  • over 2 years reverse engineering random forest predictions
  • over 2 years Model request: bam models from mgcv
  • over 2 years First call to createFolds() not reproducible... despite set.seed() call
  • over 2 years C5.0 Final Model Extract Rules / Trees
  • over 2 years categorical binning of continuous predictors
  • over 2 years Correct Way of Using update.train?