plot_effects_individual: plot_effects_individual

In livioivil/r41sqrt10: livio's sandbox

plot_effects_individual

Description

plot_effects_individual

Usage

plot_effects_individual(
  data,
  pred_name,
  resp_name,
  predict_funct,
  col.by = NULL,
  npoints = 10,
  center_effs = TRUE,
  plot_obs = TRUE
)

Arguments

data	a data.frame
pred_name	character. name of the predictor (x)
resp_name	character. name of the response (y)
predict_funct	a function that makes predictions
col.by	NULL or a vector of values that indicate the strata
npoints	10
center_effs	logical. Should be the effect of the oder predictors removed from the lines and the observations?
plot_obs	if TRUE (default) it draws the observed values (points) after residualization by the other predictors

Author(s)

livio finos

Examples

n=100
X=matrix(rnorm(n*3),n,3)
X[,2]=sign(X[,2])
y=rnorm(n,X[,1]+X[,1]*X[,2])
D=data.frame(X)
D$y=y

# Regression model
mod=lm(y~X1*X2+X3,data=D)
summary(mod)

predict_funct=function(newdata) predict(mod,newdata=newdata)
plot_effects_individual(D,"X1","y",predict_funct=predict_funct)
plot_effects_individual(D,"X1","y",predict_funct=predict_funct,col.by = D$X2)
plot_effects_individual(D,"X3","y",predict_funct=predict_funct,center_effs = TRUE)
plot_effects_individual(D,"X3","y",predict_funct=predict_funct,center_effs = FALSE)

# Regression tree model
## Not run: 
require(rpart)
mod=rpart(y~X1+X2+X3,data=D,control = list(cp=.0001))
print(mod)
printcp(mod)

predict_funct=function(newdata) predict(mod,newdata=newdata)
#plot_effects_individual(D,"X1","y",predict_funct=predict_funct)
plot_effects_individual(D,"X1","y",predict_funct=predict_funct,col.by = D$X2)
# compare the two:
plot_effects_individual(D,"X3","y",predict_funct=predict_funct,center_effs = TRUE)
plot_effects_individual(D,"X3","y",predict_funct=predict_funct,center_effs = FALSE)

## End(Not run)

livioivil/r41sqrt10 documentation built on July 29, 2023, 5:40 p.m.