langModel-class: langModel Class

In nlanderson9/languagePredictR: Predict Outcomes from Natural Language

Description

langModel Class

Slots

call

The function called to generate this model, with all arguments specified by the user

data_text

The text input to create the corpus/model

data_outcome

The outcome variable input to create the model

type

Model type, "binary" or "continuous"

text

The name of the column in the input dataframe containing the data_text

outcome

The name of the column in the input dataframe containing the data_outcome

tokens

The list of tokens in the language corpus

ngrams

The ngrams used to generate the tokens

dfmWeightScheme

The weight scheme used to create the document-frequency matrix

x

The document-frequency matrix

y

The dependent (outcome) variable

cv

The final model

lambda

The lambda value used

predicted_y

The predicted outcomes based on the model and original language data

predicted_probabilities

(If binary) The predicted probabilities of the outcomes based on the model and original language data

roc

(If binary) The ROC calculated using the predicted_y

roc_ci

(If binary) The boostrapped confidence interval calculated for the ROC

corr

(If continuous) The correlation using the predicted_y

level0

The bottom/first level of a binary variable, or the lowest value of a continuous variable

level1

The top/second level of a binary variable, or the highest value of a continuous variable

cat0raw

The predictors (word ngrams) predicting the level0 outcome, with their model weights

cat1raw

The predictors (word ngrams) predicting the level1 outcome, with their model weights

p_value

The p-value estimated via permutation test

lossMeasure

The loss measure chosen for the LASSO regression cross-validation

cvm_type

The loss measure - virtually always the same as lossMeasure except when the default "deviance" is specified

cvm

The value of the loss measure for the cross-validated model at the chosen lambda level

permuted_cvms

The value of the loss measure for each cross-validated model at the chosen lambda level across all randomized permutations

permutationK

The number of permutations conducted for permutation testing

minimum_p

The minimum p_value that can be achieved given the number permutationK specified

st_err_p

The standard error of the p_value based on the number permutationK specified

nlanderson9/languagePredictR documentation built on June 10, 2021, 11 a.m.