Feed-Forward Neural Networks and Multinomial Log-Linear Models

This learner provides feed-forward neural networks with a single hidden layer, and for multinomial log-linear models.

Format

R6Class object.

Value

Learner object with methods for both training and prediction. See Lrnr_base for documentation on learners.

Parameters

formula: A formula of the form class ~ x1 + x2 + ...
weights: (case) weights for each example – if missing defaults to 1
size: number of units in the hidden layer. Can be zero if there are skip-layer units.
entropy: switch for entropy (= maximum conditional likelihood) fitting. Default by least-squares.
decay: parameter for weight decay. Default 0.
maxit: maximum number of iterations. Default 100.
linout: switch for linear output units. Default logistic output units.
...: Other parameters passed to nnet.

Common Parameters

Individual learners have their own sets of parameters. Below is a list of shared parameters, implemented by Lrnr_base, and shared by all learners.

covariates: A character vector of covariates. The learner will use this to subset the covariates for any specified task
outcome_type: A variable_type object used to control the outcome_type used by the learner. Overrides the task outcome_type if specified
...: All other parameters should be handled by the invidual learner classes. See the documentation for the learner class you're instantiating

Other Learners: Custom_chain, Lrnr_HarmonicReg, Lrnr_arima, Lrnr_bartMachine, Lrnr_base, Lrnr_bayesglm, Lrnr_bilstm, Lrnr_caret, Lrnr_cv_selector, Lrnr_cv, Lrnr_dbarts, Lrnr_define_interactions, Lrnr_density_discretize, Lrnr_density_hse, Lrnr_density_semiparametric, Lrnr_earth, Lrnr_expSmooth, Lrnr_gam, Lrnr_ga, Lrnr_gbm, Lrnr_glm_fast, Lrnr_glm_semiparametric, Lrnr_glmnet, Lrnr_glmtree, Lrnr_glm, Lrnr_grfcate, Lrnr_grf, Lrnr_gru_keras, Lrnr_gts, Lrnr_h2o_grid, Lrnr_hal9001, Lrnr_haldensify, Lrnr_hts, Lrnr_independent_binomial, Lrnr_lightgbm, Lrnr_lstm_keras, Lrnr_mean, Lrnr_multiple_ts, Lrnr_multivariate, Lrnr_nnls, Lrnr_optim, Lrnr_pca, Lrnr_pkg_SuperLearner, Lrnr_polspline, Lrnr_pooled_hazards, Lrnr_randomForest, Lrnr_ranger, Lrnr_revere_task, Lrnr_rpart, Lrnr_rugarch, Lrnr_screener_augment, Lrnr_screener_coefs, Lrnr_screener_correlation, Lrnr_screener_importance, Lrnr_sl, Lrnr_solnp_density, Lrnr_solnp, Lrnr_stratified, Lrnr_subset_covariates, Lrnr_svm, Lrnr_tsDyn, Lrnr_ts_weights, Lrnr_xgboost, Pipeline, Stack, define_h2o_X(), undocumented_learner

Examples

set.seed(123)

# load example data
data(cpp_imputed)
covars <- c("bmi", "parity", "mage", "sexn")
outcome <- "haz"

# create sl3 task
task <- sl3_Task$new(cpp_imputed, covariates = covars, outcome = outcome)

# train neural networks and make predictions
lrnr_nnet <- Lrnr_nnet$new(linout = TRUE, size = 10, maxit = 1000)
fit <- lrnr_nnet$train(task)
#> # weights:  61
#> initial  value 3473.327147 
#> iter  10 value 2332.708133
#> iter  20 value 2285.802094
#> iter  30 value 2259.282969
#> iter  40 value 2186.037648
#> iter  50 value 2163.163175
#> iter  60 value 2158.274072
#> iter  70 value 2157.252818
#> iter  80 value 2153.548324
#> iter  90 value 2147.689763
#> iter 100 value 2140.508801
#> iter 110 value 2136.151335
#> iter 120 value 2131.667979
#> iter 130 value 2127.812539
#> iter 140 value 2126.292145
#> iter 150 value 2121.169605
#> iter 160 value 2119.921691
#> iter 170 value 2117.912191
#> iter 180 value 2116.976429
#> iter 190 value 2110.064796
#> iter 200 value 2105.265439
#> iter 210 value 2090.487599
#> iter 220 value 2083.288260
#> iter 230 value 2078.431944
#> iter 240 value 2075.836591
#> iter 250 value 2072.891518
#> iter 260 value 2069.692797
#> iter 270 value 2067.111192
#> iter 280 value 2065.639707
#> iter 290 value 2064.105394
#> iter 300 value 2059.875667
#> iter 310 value 2059.346266
#> iter 320 value 2059.315552
#> iter 330 value 2059.104485
#> final  value 2059.103001 
#> converged
preds <- fit$predict(task)