Man pages for DexGroves/hacktoolkit
Generic Hackathon Sauce
| accuracy | Accuracy. |
| add_encoding | Add an interaction, poly, one-hot style effect that can be... |
| auc | Classic, binomial AUC. |
| binomial_deviance | Calculate binomial deviance. |
| cross_train_predict | Train and predict out-of-fold so that predictions can be used... |
| h2o_cvdf | Parse the h2o cv grid into something less ridiculous. |
| h2o_plotcv | Plot the CV error vis-a-vis plot.cv.glmnet. Except without... |
| logit_to_prob | Apply logit transform on a vector x. |
| loo_encode | Do Owen Zhang style leave-one-out encoding of a categorical. |
| marginal_effect | Get pseudo marginal effects from any model. |
| model_conform | Form a model matrix with structure forced to correspond to... |
| multivariate_auc | Multivariate AUC. |
| mv_binomial_deviance | Multivariate binomial deviance. |
| predict_folds | Get a matrix of predictions for a list of models. Parallelise... |
| prob_to_logit | Apply the inverse logit transform on a vector x. |
| train_folds | Train crossfold models. Parallelise with registerDoMC. |
| weighted_var | Calculate weighted variance. |
| xgb_perf | Imitate gbm::gbm.perf for an xgboost::xgb.cv object. |
