Package 'predict3d'

Title: Draw Three Dimensional Predict Plot Using Package 'rgl'
Description: Draw 2 dimensional and three dimensional plot for multiple regression models using package 'ggplot2' and 'rgl'. Supports linear models (lm), generalized linear models (glm) and local polynomial regression fittings (loess).
Authors: Keon-Woong Moon [aut, cre]
Maintainer: Keon-Woong Moon <[email protected]>
License: GPL-2
Version: 0.1.5
Built: 2025-01-30 02:51:58 UTC
Source: https://github.com/cardiomoon/predict3d

Help Index

Add lines with labels to pre-existing ggplot

Description

Add lines with labels to pre-existing ggplot

Usage

add_lines(
  p,
  df,
  xpos = 0.3,
  add.coord.fixed = TRUE,
  lty = NULL,
  color = NULL,
  size = 0.5,
  add_theme_bw2 = TRUE,
  ...
)

Arguments

p

An object of class ggplot
df

A data.frame. Required columns are slope, intercept and label
xpos

A numeric. Relative horizontal position
add.coord.fixed

Logical. Whether or not add coord_fixed() function
lty

line type
color

line color
size

line size
add_theme_bw2

logical Whether or not add theme_bw2()
...

Further arguments to be passed to geom_text

Examples

require(ggplot2)
fit=lm(mpg~wt*hp,data=mtcars)
df=calEquation(fit)
p=ggplot(data=mtcars,aes(x=wt,y=mpg))
add_lines(p,df)
add_lines(p,df,lty=1:3,color=1:3,size=1)
fit=lm(mpg~wt*vs,data=mtcars)
df=calEquation(fit)
p=ggplot(data=mtcars)+geom_point(aes(x=wt,y=mpg))
add_lines(p,df)
add_lines(p,df,lty=1:2,color=1:2,size=1)+theme_bw()

Whether a string vector can be converted to numeric

Description

Whether a string vector can be converted to numeric

Usage

beNumeric(x)

Arguments

x

A string vector

Examples

x=c("age","22.5","11/2")
beNumeric(x)

calculated slope and intercept from object of class lm

Description

calculated slope and intercept from object of class lm

Usage

calEquation(
  fit,
  mode = 1,
  pred = NULL,
  modx = NULL,
  modx.values = NULL,
  label = NULL,
  maxylev = 6,
  digits = 2
)

Arguments

fit

An object of class lm
mode

A numeric
pred

name of predictor variable
modx

name of modifier variable
modx.values

Numeric. Values of modifier variable
label

A character string
maxylev

maximum length of unique value of variable to be treated as a categorial variable
digits

Integer indicating the number of decimal places

Examples

fit=lm(mpg~wt*hp+carb,data=mtcars)
calEquation(fit)
calEquation(fit,pred="hp")

expand.grid with two data.frames

Description

expand.grid with two data.frames

Usage

expand.grid2(df1, df2)

Arguments

df1

A data.frame
df2

A data.frame

Make a new data set for prediction

Description

Make a new data set for prediction

Usage

fit2newdata(
  fit,
  predictors,
  mode = 1,
  pred.values = NULL,
  modx.values = NULL,
  mod2.values = NULL,
  colorn = 3,
  maxylev = 6,
  summarymode = 1
)

Arguments

fit

An object of class "lm", "glm" or "loess"
predictors

Names of predictor variables in string
mode

A numeric. Useful when the variables are numeric. If 1, c(-1,0,1)*sd + mean is used. If 2, the 16th, 50th, 84th percentile values used. If 3 sequence over a the range of a vector used
pred.values

For which values of the predictors should be used? Default is NULL. If NULL, 20 seq_range is used.
modx.values

For which values of the moderator should lines be plotted? Default is NULL. If NULL, then the customary +/- 1 standard deviation from the mean as well as the mean itself are used for continuous moderators. If the moderator is a factor variable and modx.values is NULL, each level of the factor is included.
mod2.values

For which values of the second moderator should lines be plotted? Default is NULL. If NULL, then the customary +/- 1 standard deviation from the mean as well as the mean itself are used for continuous moderators. If the moderator is a factor variable and modx.values is NULL, each level of the factor is included.
colorn

The number of regression lines when the modifier variable(s) are numeric.
maxylev

An integer indicating the maximum number of levels of numeric variable be treated as a categorical variable
summarymode

An integer indicating method of extracting typical value of variables. If 1, typical() is used.If 2, mean() is used.

Examples

fit=lm(mpg~hp*wt*cyl+carb+am,data=mtcars)
fit2newdata(fit,predictors=c("hp","wt","am"))
fit2newdata(fit,predictors=c("hp","wt","cyl"))
fit2newdata(fit,predictors=c("hp"))
fit2newdata(fit,predictors=c("hp","wt"))
fit=loess(mpg~hp*wt*am,data=mtcars)
fit2newdata(fit,predictors=c("hp"))
## Not run: 
mtcars$engine=ifelse(mtcars$vs==0,"V-shaped","straight")
fit=lm(mpg~wt*engine,data=mtcars)
fit2newdata(fit,predictors=c("wt","engine"))
fit=lm(mpg~wt*factor(vs),data=mtcars)
fit2newdata(fit,predictors=c("wt","vs"))
fit2newdata(lm(mpg~hp*wt,data=mtcars),predictors=c("hp","wt"),mode=3,colorn=30)
fit=lm(mpg~hp*log(wt),data=mtcars)
fit2newdata(fit,predictors=c("hp","log(wt)"))
fit=lm(mpg~hp*wt*factor(vs),data=mtcars)
fit2newdata(fit,predictors=c("hp"))

## End(Not run)
require(moonBook)
fit=lm(log(NTAV)~I(age^2)*sex,data=radial)
fit2newdata(fit,predictors=c("I(age^2)","sex"))

Get aspect information of a ggplot

Description

Get aspect information of a ggplot

Usage

getAspectRatio(p)

Arguments

p

A ggplot object

calculate mean values of two consecutive number

Description

calculate mean values of two consecutive number

Usage

getMeans(x)

Arguments

x

A numeric vector

Examples

x=c(50,60,70)
getMeans(x)

Make new formula

Description

Make new formula

Usage

getNewFormula(fit, predictors = NULL)

Arguments

fit

An object of class lm or glm
predictors

Names of variables to exclude

Examples

fit=lm(mpg~factor(cyl)*factor(am)+wt+carb,data=mtcars)
getNewFormula(fit,predictors=c("cyl","wt"))
fit=lm(Sepal.Length~Sepal.Width*Petal.Length+Species,data=iris)
getNewFormula(fit,predictors=c("Petal.Length"))
fit=lm(mpg~hp*wt*factor(cyl),data=mtcars)
getNewFormula(fit,predictors=c("hp","cyl"))
fit=loess(mpg~hp*wt,data=mtcars)
getNewFormula(fit,predictors=c("hp","wt"))

Pick default color

Description

Pick default color

Usage

gg_color_hue(n)

Arguments

n

An integer

Visualize predictions from the multiple regression models.

Description

Visualize predictions from the multiple regression models.

Usage

ggPredict(
  fit,
  pred = NULL,
  modx = NULL,
  mod2 = NULL,
  modx.values = NULL,
  mod2.values = NULL,
  dep = NULL,
  mode = 1,
  colorn = 3,
  maxylev = 6,
  show.point = getOption("ggPredict.show.point", TRUE),
  show.error = FALSE,
  error.color = "red",
  jitter = NULL,
  se = FALSE,
  alpha = 0.1,
  show.text = TRUE,
  add.modx.values = TRUE,
  add.loess = FALSE,
  labels = NULL,
  angle = NULL,
  xpos = NULL,
  vjust = NULL,
  digits = 2,
  facet.modx = FALSE,
  facetbycol = TRUE,
  plot = TRUE,
  summarymode = 1,
  ...
)

Arguments

fit

An object of class "lm" or "glm"
pred

The name of predictor variable
modx

Optional. The name of moderator variable
mod2

Optional. The name of second moderator variable
modx.values

For which values of the moderator should lines be plotted? Default is NULL. If NULL, then the customary +/- 1 standard deviation from the mean as well as the mean itself are used for continuous moderators. If the moderator is a factor variable and modx.values is NULL, each level of the factor is included.
mod2.values

For which values of the second moderator should lines be plotted? Default is NULL. If NULL, then the customary +/- 1 standard deviation from the mean as well as the mean itself are used for continuous moderators. If the moderator is a factor variable and modx.values is NULL, each level of the factor is included.
dep

Optional. The name of dependent variable
mode

A numeric. Useful when the variables are numeric. If 1, c(-1,0,1)*sd + mean is used. If 2, the 14th, 50th, 86th percentile values used. If 3 sequence over a the range of a vector used
colorn

The number of regression lines when the modifier variable(s) are numeric.
maxylev

An integer indicating the maximum number of levels of numeric variable be treated as a categorical variable
show.point

Logical. Whether or not add points
show.error

Logical. Whether or not show error
error.color

color of error. dafault value is "red"
jitter

logical Whether or not use geom_jitter
se

Logical. Whether or not add confidence interval
alpha

A numeric. Transparency
show.text

Logical. Whether or not add regression equation as label
add.modx.values

Logical. Whether or not add moderator values to regression equation
add.loess

Logical. Whether or not add loess line
labels

labels on regression lines
angle

angle of text
xpos

x axis position of label
vjust

vertical alignment of labels
digits

integer indicating the number of decimal places
facet.modx

Create separate panels for each level of the moderator? Default is FALSE
facetbycol

Logical.
plot

Logical. Should a plot of the results be printed? Default is TRUE.
summarymode

An integer indicating method of extracting typical value of variables. If 1, typical() is used.If 2, mean() is used.
...

additional arguments to be passed to geom_text

Examples

fit=loess(mpg~hp*wt*am,data=mtcars)
ggPredict(fit)
ggPredict(fit,hp)
## Not run: 
ggPredict(fit,hp,wt)
fit=lm(mpg~wt*hp-1,data=mtcars)
ggPredict(fit,xpos=0.7)
fit=lm(mpg~hp*wt,data=mtcars)
ggPredict(fit)
ggPredict(fit,labels=paste0("label",1:3),xpos=c(0.3,0.6,0.4))
ggPredict(fit,se=TRUE)
ggPredict(fit,mode=3,colorn=40,show.text=FALSE)
fit=lm(log(mpg)~hp*wt,data=mtcars)
ggPredict(fit,dep=mpg)
fit=lm(mpg~hp*wt*cyl,data=mtcars)
ggPredict(fit,modx=wt,modx.values=c(2,3,4,5),mod2=cyl,show.text=FALSE)
ggPredict(fit,hp,wt,show.point=FALSE,se=TRUE,xpos=0.5)
ggPredict(fit,modx=wt,xpos=0.3)
ggPredict(fit)
mtcars$engine=ifelse(mtcars$vs==0,"V-shaped","straight")
fit=lm(mpg~wt*engine,data=mtcars)
ggPredict(fit)
require(TH.data)
fit=glm(cens~pnodes*horTh,data=GBSG2,family=binomial)
ggPredict(fit,pnodes,horTh,se=TRUE,xpos=c(0.6,0.3),angle=c(40,60),vjust=c(2,-0.5))
fit1=glm(cens~pnodes,data=GBSG2,family=binomial)
ggPredict(fit1,vjust=1.5,angle=45)
fit3=glm(cens~pnodes*age,data=GBSG2,family=binomial)
ggPredict(fit3,pred=pnodes,modx=age,mode=3,colorn=10,show.text=FALSE)
fit2=glm(cens~pnodes*age*horTh,data=GBSG2,family=binomial)
ggPredict(fit2,pred=pnodes,modx=age,mod2=horTh,mode=3,colorn=10,show.text=FALSE)
fit=lm(mpg~log(hp)*wt,data=mtcars)
ggPredict(fit,hp,wt)
fit=lm(mpg~hp*wt+disp+gear+carb+am,data=mtcars)
ggPredict(fit,disp,gear,am)
library(moonBook)
fit=lm(weight~I(height^3)+I(height^2)+height+sex,data=radial)
ggPredict(fit)
predict3d(fit)

## End(Not run)

Decide whether a vector can be treated as a numeric variable

Description

Decide whether a vector can be treated as a numeric variable

Usage

is.mynumeric(x, maxylev = 6)

Arguments

x

A vector
maxylev

An integer indicating the maximum number of levels of numeric variable be treated as a categorical variable

Generate regular sequences of desired length between minimum and maximal values

Description

Generate regular sequences of desired length between minimum and maximal values

Usage

myseq(x, length = 20)

Arguments

x

a numeric vector
length

desired length of the sequence

Convert a numeric vector into groups

Description

Convert a numeric vector into groups

Usage

number2group(
  x,
  mode = 1,
  values = NULL,
  silent = FALSE,
  label = "label",
  digits = 2,
  colorn = 3
)

Arguments

x

A numeric vector
mode

A numeric. If 1, mean(x) +c(-1,0,1)*sd(x) are used. If 2, quantile(x,probs=c(0.14,0.5,0.86),type=6) are used. If 3, values are used
values

A numeric vector
silent

A logical. Whether table of result will be shown
label

A character string
digits

integer indicating the number of decimal places
colorn

The number of regression lines when the modifier variable(s) are numeric

Examples

number2group(iris$Sepal.Length,label="Sepal.Length")
x=number2group(mtcars$wt,label="wt")
x

Draw 3d predict plot using package 'rgl'

Description

Draw 3d predict plot using package 'rgl'

Usage

predict3d(
  fit,
  pred = NULL,
  modx = NULL,
  mod2 = NULL,
  dep = NULL,
  xlab = NULL,
  ylab = NULL,
  zlab = NULL,
  width = 640,
  colorn = 20,
  maxylev = 6,
  se = FALSE,
  show.summary = FALSE,
  overlay = NULL,
  show.error = FALSE,
  show.legend = FALSE,
  bg = NULL,
  type = "s",
  radius = NULL,
  palette = NULL,
  palette.reverse = TRUE,
  color = "red",
  show.subtitle = TRUE,
  show.plane = TRUE,
  plane.color = "steelblue",
  plane.alpha = 0.5,
  summarymode = 1,
  ...
)

Arguments

fit

A model object for which prediction is desired.
pred

The name of predictor variable
modx

Optional. The name of moderator variable
mod2

Optional. The name of second moderator variable
dep

Optional. The name of dependent variable
xlab

x-axis label.
ylab

y-axis label.
zlab

z-axis label.
width

the width of device
colorn

An integer giving the desired number of intervals. Non-integer values are rounded down.
maxylev

Maximal length of unique values of y axis variable to be treated as a categorical variable.
se

Logical. Whether or not show se. Only effective when the y-axis variable is a categorical one.
show.summary

Logical. Whether or not show statistical summary
overlay

Logical. Whether or not overlay plots
show.error

Logical. Whether or not show error
show.legend

Logical. Whether or not show legend
bg

Character. Background color of plot
type

For the default method, a single character indicating the type of item to plot. Supported types are: 'p' for points, 's' for spheres, 'l' for lines, 'h' for line segments from z = 0, and 'n' for nothing. For the mesh3d method, one of 'shade', 'wire', or 'dots'. Partial matching is used.
radius

The size of sphere
palette

Name of color palette
palette.reverse

Logical. Whether or not reverse the palette order
color

Default color. Color is used when the palette is NULL
show.subtitle

Logical. If true, show regression call as subtitle
show.plane

Logical. If true, show regression plane
plane.color

Name of color of regression plane
plane.alpha

Transparency scale of regression plane
summarymode

An integer indicating method of extracting typical value of variables. If 1, typical() is used.If 2, mean() is used.
...

additional parameters which will be passed to plot3d

Examples

fit=lm(mpg~hp*wt,data=mtcars)
predict3d(fit,show.error=TRUE)
fit=lm(log(mpg)~hp*wt,data=mtcars)
predict3d(fit,dep=mpg)
## Not run: 
fit=lm(Sepal.Length~Sepal.Width*Species,data=iris)
predict3d(fit)
require(TH.data)
fit=glm(cens~pnodes*age*horTh,data=GBSG2,family=binomial)
predict3d(fit)
mtcars$engine=ifelse(mtcars$vs==0,"V-shaped","straight")
fit=lm(mpg~wt*engine,data=mtcars)
predict3d(fit)
fit=loess(mpg~hp*wt,data=mtcars)
predict3d(fit,radius=4)
states<-as.data.frame(state.x77[,c("Murder","Population","Illiteracy","Income","Frost")])
fit=lm(Murder~Population+Illiteracy,data=states)
predict3d(fit)
predict3d(fit,radius=200)
fit=lm(mpg~cyl+hp+am,data=mtcars)
predict3d(fit)

## End(Not run)

Rank a numeric vector using proportional table and returns character vector of names of color using palette

Description

Rank a numeric vector using proportional table and returns character vector of names of color using palette

Usage

rank2colors(x, palette = "Blues", reverse = TRUE, color = "red")

Arguments

x

A numeric vector
palette

Name of the color palette
reverse

Logical. Whether or not reverse the order of the color palette
color

Default color when palette is NULL

Examples

rank2colors(mtcars$wt,palette="Blues")

Rank a numeric vector using proportional table and returns a new ordinal vector

Description

Rank a numeric vector using proportional table and returns a new ordinal vector

Usage

rank2group2(x, k = 4)

Arguments

x

a numeric vector
k

a integer specifies how many groups you want to classify. default value is 4

Restore factors in data.frame as numeric

Description

Restore factors in data.frame as numeric

Usage

restoreData(data)

Arguments

data

A data.frame

Examples

fit=lm(mpg~factor(cyl)*factor(am),data=mtcars)
fit=lm(mpg~wt*factor(am),data=mtcars)
fit=lm(mpg~wt*hp,data=mtcars)
restoreData(fit$model)

restore data column with I() function

Description

restore data column with I() function

Usage

restoreData2(df)

Arguments

df

A data.frame

Examples

fit=lm(mpg~I(cyl^(1/2))*am,data=mtcars)
restoreData2(fit$model)
fit=lm(mpg~sqrt(hp)*log(wt)*am,data=mtcars)
restoreData2(fit$model)

Restore data from arithmetic operator

Description

Restore data from arithmetic operator

Usage

restoreData3(df, changeLabel = FALSE)

Arguments

df

A data.frame
changeLabel

logical

Examples

fit=lm(2^mpg~hp*wt,data=mtcars)
summary(fit)
restoreData3(fit$model)

Restore factors in variable name as numeric

Description

Restore factors in variable name as numeric

Usage

restoreNames(x)

Arguments

x

character vector

Examples

restoreNames(c("factor(cyl)","am"))
restoreNames(c("I(age^2)","am","100/mpg","cyl^1/2","mpg2","sex + 0.5"))

get opposite arithmetic operator

Description

get opposite arithmetic operator

Usage

revOperator(operator)

Arguments

operator

A character

Find variable names in data.frame

Description

Find variable names in data.frame

Usage

seekNamesDf(vars, df)

Arguments

vars

variable names to find
df

A data.frame

Value

A character vector

Create a sequence over the range of a vector

Description

Create a sequence over the range of a vector

Usage

seq_range(x, n = 2)

Arguments

x

A numeric vector
n

An integer specifying the length of sequence (i.e., number of points across the range of x)

Examples

seq_range(1:5,n=3)

Make angle data with slope data

Description

Make angle data with slope data

Usage

slope2angle(
  df,
  fit,
  ytransform = 0,
  predc,
  temppredc,
  modxc,
  yvar,
  p,
  method = "lm",
  xpos = NULL,
  vjust = NULL,
  digits = 3,
  facetno = NULL,
  add.modx.values = TRUE
)

Arguments

df

A data.frame
fit

An object of class "lm" or "glm"
ytransform

Numeric. If 1, log transformation of dependent variable, If -1, exponential transformation
predc

Name of predictor variable
temppredc

Name of predictor variable in regression equation
modxc

Name of moderator variable
yvar

Name of dependent variable
p

An object of class ggplot
method

String. Choices are one of "lm" and "glm".
xpos

The relative x-axis position of labels. Should be within 0 to 1
vjust

vjust
digits

integer indicating the number of decimal places
facetno

The number of facets
add.modx.values

Whether add name of moderator variable

change string to pattern

Description

change string to pattern

Usage

string2pattern(string)

Arguments

string

A character vector

Examples

string=c("I(age^2)","factor(cyl)","log(mpg)")
string2pattern(string)

theme_bw with no grid

Description

theme_bw with no grid

Usage

theme_bw2()