prepare_xgboost_model: NLP modeling using xgboost.

In tomathon-io/wordly: Natural Language Processing (NLP) functionality and modeling

Description

Allows for modeling of Natural Language Processing (NLP) data using the xgboost package. Default param values are optimized for NLP.

Usage

prepare_xgboost_model(train_data, dtm_train_data, response_label_name,
  xgb_objective, xgb_nrounds = 100, xgb_max_depth = 20,
  xgb_eta = 0.1, xgb_nthread = 4, xgb_weight = NULL,
  xgb_missing = NA)

Arguments

train_data	The NON-DTM training input data.
dtm_train_data	The DTM training input data. Returned from prepare_dtm().
response_label_name	The name of the column in dat_in containing the text source.
xgb_objective	The xgboost parameter for objective (see xgboost documentation). One of "reg:linear", "binary:logistic", etc.
xgb_nrounds	The xgboost parameter for nrounds (see xgboost documentation). Defaults to 100.
xgb_max_depth	The xgboost parameter for max_depth (see xgboost documentation). Defaults to 20.
xgb_eta	The xgboost parameter for eta (see xgboost documentation). Defaults to 0.1.
xgb_nthread	The xgboost parameter for nthreads (see xgboost documentation). Defaults to 4.
xgb_weight	The xgboost parameter for weight (see xgboost documentation). Defaults to NULL.
xgb_missing	The xgboost parameter for missing (see xgboost documentation). Defaults to NA.

Examples

xgb_mod <- prepare_xgboost_model(train_, dtm_train, "product_review", xgb_objective = "binary:logistic")

tomathon-io/wordly documentation built on June 15, 2020, 12:41 a.m.