importance.test.boosting {logicDT}

Term importance test based on boosted linear models

Description

This function takes a fitted linear.logic model and independent test data as input for testing if the included terms are influential with respect to the outcome. This hypothesis test is based on a likelihood-ratio test.

Usage

importance.test.boosting(model, X, y, Z, Z.interactions = TRUE)

Arguments

model

A fitted linear.logic model (i.e., a model created via fitLinearLogicModel or fitLinearBoostingModel)

X

Matrix or data frame of binary input data. This object should correspond to the binary matrix for fitting the model.

y

Response vector. 0-1 coding for binary outcomes.

Z

Optional quantitative covariables supplied as a matrix or data frame. Only used (and required) if the model was fitted using them.

Z.interactions

A Boolean value determining whether interactions with quantitative covaraible Z shall be taken into account

Details

In detail, the null hypotheses

H_0: \beta_j = \delta_j = 0

using the linear model

g(E[Y]) = \beta_0 + \sum_{i=1}^B \beta_i \cdot 1[C_i] + \delta_0 \cdot E + \sum_{i=1}^B \delta_i \cdot 1[C_i] \cdot E

are tested for each j \in \lbrace 1,\ldots,B \rbrace if Z.interactions is set to TRUE. Otherwise, the null hypotheses

H_0: \beta_j = 0

using the linear model

g(E[Y]) = \beta_0 + \sum_{i=1}^B \beta_i \cdot 1[C_i] + \delta_0 \cdot E

are tested.

Value

A data frame consisting of three columns,

var

The tested term,

vim

The associated variable importance, and

p.value

The corresponding p-value for testing if the term is influential.