gpanel.band: Panel functions for predicted values and SE bands
In gmonette/spida15: Collection of miscellaneous functions for mixed models etc. prepared for SPIDA 2009+ (development version)
Description
Usage
Arguments
Details
Value
Author(s)
Examples
Description
Panel functions for predicted values and SE bands using 'layer' and 'glayer' in the package latticeExtra
Usage
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
panel.fit(x, y, fit,
subscripts, ..., type = 'l', font, fontface)
panel.fit(...) # when called within 'layer' or 'glayer' in 'latticeExtra'
panel.band(x, y, lower, upper, subscripts, ...,
col, group.number, alpha,
col.symbol, border = F, font, fontface) # see examples belor
panel.band(...) # when called within 'layer' or 'glayer' in 'latticeExtra'
panel.labels(x, y, labels , subscripts, ...)
panel.labels(...) # when called within 'layer' or 'glayer' in 'latticeExtra'
fillin(data, form, n = 200, xpd = 1.0)
Arguments
data
data frame to be used to add additional values of numeric variable
form
formula evaluated in data. The first term defines the variable with values to be filled in and the remaining terms define the variables to be used for grouping determining the minima and maxima within which values are added.
n
the number of values to be added between the global mininum and maximum. Values falling outside conditional minima and maxima are culled. Default 200.
xpd
expansion factor to add points beyond minima and maxima. Default 1.0.
y
fit
fitted values of a model, generally passed through 'layer' from a call to 'xyplot': e.g. xyplot( y ~ x, data, groups = g, fit = data$yhat, lower = with(data, yhat - 2*se), upper = with(data, yhat + 2*se), subscripts = T)
lower
upper
subscripts
...
col
group.number
alpha
to deal with the problem that 'trellis.par.get' will pass a value of alpha = 1, 'panel.band' will set alpha = .3 if alpha = 1. To get a nearly opaque band, use alpha = .99.
col.symbol
is used to control color when using 'groups'
border
default = FALSE for panel.band.
font
fontface
Details
With 'layer' and 'glayer' in 'latticeExtra', these functions can be used to easily generate fitted values and confidence or prediction bands that have a reasonable appearance whether a plot uses 'groups' or not.
Value
The 'panel.bands', 'panel.fit', and 'panel.labels' functions are invoked for their graphical effect.
Author(s)
Georges Monette <georges@yorku.ca>
Examples
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
##---- Should be DIRECTLY executable !! ----
##-- ==> Define data, use random,
##-- or do help(data=index) for the standard data sets.
library(spida)
library(latticeExtra)
library(car)
fit <- lm(prestige ~ (income + I(income^2)) * type, Prestige,
na.action = na.exclude)
pred <- cbind( Prestige, predict(fit, newdata = Prestige, se = TRUE))
head(pred)
(p <- xyplot( prestige ~ income , pred, groups = type,
subscripts = T,
fit = pred$fit,
lower = with(pred, fit - 2*se.fit),
upper = with(pred, fit + 2*se.fit)))
p + glayer(gpanel.band(...))
p + glayer(gpanel.band(...)) + glayer(gpanel.fit(...))
# Another example
# to illustrate 'fillin'
fit <- lm( income ~
(education+I(education^2)+I(education^3) +I(education^4))* type,
Prestige, na.action = na.exclude) # overfitting!
# adding extra values of predictor to get smooth line
Prestige$occupation <- rownames(Prestige)
z <- fillin(Prestige, ~education + type,xpd = 1.1)
dim(z)
dim(Prestige)
z <- cbind(z, predict(fit, newdata = z, se = TRUE))
head(z)
gd(3,cex=2,lwd=2, alpha = .7)
(p <- xyplot( income ~ education, z, groups = type,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
auto.key = list(space='right', lines= T)))
p + glayer( gpanel.fit(...))
p + glayer( gpanel.fit(...)) + glayer(gpanel.band(...))
p + glayer( gpanel.band(..., alpha = .1))
gd(3,lty=1,lwd=2)
p + glayer( gpanel.band(...,alpha = .1)) + glayer(gpanel.fit(...))
# with panels and no groups:
(p <- xyplot( income ~ education| type, z,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
auto.key = list(space='right', lines= T)))
p + layer( gpanel.fit(...)) + layer( gpanel.band(...))
gd(basecol = 'tomato4')
p + layer( gpanel.band(..., col = 'grey10')) + layer(gpanel.fit(...))
p + layer( gpanel.band(...)) + layer(gpanel.fit(...))
# With panels and groups
z <- Prestige
z$gender <- with(z, cut( women, c(-1,15,50,101),labels = c("Male","Mixed","Female")))
tab(z, ~ gender + type)
z <- fillin( z, ~ education + type + gender, xpd = 1.1)
fit <- lm( income ~ (education+I(education^2)+I(education^3) )* type * gender,
z, na.action = na.exclude) # overfitting!
summary(fit)
z <- cbind( z, predict(fit, newdata = z, se = TRUE))
head(z)
(p <- xyplot( income ~ education| gender, z, groups = type,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
layout = c(1,3),
auto.key = list(space='right', lines= T, cex = 1.5)))
p + glayer( gpanel.band(...)) + glayer(gpanel.fit(...))
# trellis.focus()
# panel.identify(labels= z$occupation)
# trellis.unfocus()
z$type2 <- with( z, reorder(type,education, mean, na.rm=T))
gd(3)
(p <- xyplot( income ~ education| type2, z, groups = gender,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
layout = c(1,3),
par.strip.text = list(cex = 2),
auto.key = list(space='right', lines= T, cex = 1.5)))
p + glayer( gpanel.fit(...))
p + glayer( gpanel.band(...))
p + glayer( gpanel.band(...)) + glayer(gpanel.fit(...))
# trellis.focus()
# panel.identify(labels= z$occupation)
# trellis.unfocus()
# With panels^2
# need to remove 'col = col.line'
z <- Prestige
z$occ <- rownames(Prestige)
z$gender <- with(z, cut( women, c(-1,15,50,101),labels = c("Male","Mixed","Female")))
z$type2 <- with( z, reorder(type,education, mean, na.rm=T))
tab(z, ~ gender + type2)
z <- fillin( z, ~ education + type + gender, xpd = 1.1)
fit <- lm( income ~ (education+I(education^2)+I(education^3) )* type * gender,
z, na.action = na.exclude) # overfitting!
summary(fit)
z <- cbind( z, predict(fit, newdata = z, se = TRUE))
head(z)
(p <- xyplot( income ~ education| gender*type, z,
subscripts = T,
fit = z$fit,
labels = z$occ,
lower = z$fit - z$se,
upper = z$fit + z$se,
auto.key = list(space='right', lines= T, cex = 1.5)))
p + layer( gpanel.fit(...)) + layer( gpanel.band(...))
p + layer( gpanel.band(..., col = 'black', alpha = .1)) + layer(gpanel.fit(...)) +
layer(gpanel.text(...))
gmonette/spida15 documentation built on May 17, 2019, 7:26 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Description Usage Arguments Details Value Author(s) Examples
Description
Panel functions for predicted values and SE bands using 'layer' and 'glayer' in the package latticeExtra
Usage
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | panel.fit(x, y, fit,
subscripts, ..., type = 'l', font, fontface)
panel.fit(...) # when called within 'layer' or 'glayer' in 'latticeExtra'
panel.band(x, y, lower, upper, subscripts, ...,
col, group.number, alpha,
col.symbol, border = F, font, fontface) # see examples belor
panel.band(...) # when called within 'layer' or 'glayer' in 'latticeExtra'
panel.labels(x, y, labels , subscripts, ...)
panel.labels(...) # when called within 'layer' or 'glayer' in 'latticeExtra'
fillin(data, form, n = 200, xpd = 1.0)
|
Arguments
data |
data frame to be used to add additional values of numeric variable |
form |
formula evaluated in data. The first term defines the variable with values to be filled in and the remaining terms define the variables to be used for grouping determining the minima and maxima within which values are added. |
n |
the number of values to be added between the global mininum and maximum. Values falling outside conditional minima and maxima are culled. Default 200. |
xpd |
expansion factor to add points beyond minima and maxima. Default 1.0. |
y |
|
fit |
fitted values of a model, generally passed through 'layer' from a call to 'xyplot': e.g. |
lower |
|
upper |
|
subscripts |
|
... |
|
col |
|
group.number |
|
alpha |
to deal with the problem that 'trellis.par.get' will pass a value of alpha = 1, 'panel.band' will set alpha = .3 if alpha = 1. To get a nearly opaque band, use alpha = .99. |
col.symbol |
is used to control color when using 'groups' |
border |
default = FALSE for panel.band. |
font |
|
fontface |
Details
With 'layer' and 'glayer' in 'latticeExtra', these functions can be used to easily generate fitted values and confidence or prediction bands that have a reasonable appearance whether a plot uses 'groups' or not.
Value
The 'panel.bands', 'panel.fit', and 'panel.labels' functions are invoked for their graphical effect.
Author(s)
Georges Monette <georges@yorku.ca>
Examples
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 | ##---- Should be DIRECTLY executable !! ----
##-- ==> Define data, use random,
##-- or do help(data=index) for the standard data sets.
library(spida)
library(latticeExtra)
library(car)
fit <- lm(prestige ~ (income + I(income^2)) * type, Prestige,
na.action = na.exclude)
pred <- cbind( Prestige, predict(fit, newdata = Prestige, se = TRUE))
head(pred)
(p <- xyplot( prestige ~ income , pred, groups = type,
subscripts = T,
fit = pred$fit,
lower = with(pred, fit - 2*se.fit),
upper = with(pred, fit + 2*se.fit)))
p + glayer(gpanel.band(...))
p + glayer(gpanel.band(...)) + glayer(gpanel.fit(...))
# Another example
# to illustrate 'fillin'
fit <- lm( income ~
(education+I(education^2)+I(education^3) +I(education^4))* type,
Prestige, na.action = na.exclude) # overfitting!
# adding extra values of predictor to get smooth line
Prestige$occupation <- rownames(Prestige)
z <- fillin(Prestige, ~education + type,xpd = 1.1)
dim(z)
dim(Prestige)
z <- cbind(z, predict(fit, newdata = z, se = TRUE))
head(z)
gd(3,cex=2,lwd=2, alpha = .7)
(p <- xyplot( income ~ education, z, groups = type,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
auto.key = list(space='right', lines= T)))
p + glayer( gpanel.fit(...))
p + glayer( gpanel.fit(...)) + glayer(gpanel.band(...))
p + glayer( gpanel.band(..., alpha = .1))
gd(3,lty=1,lwd=2)
p + glayer( gpanel.band(...,alpha = .1)) + glayer(gpanel.fit(...))
# with panels and no groups:
(p <- xyplot( income ~ education| type, z,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
auto.key = list(space='right', lines= T)))
p + layer( gpanel.fit(...)) + layer( gpanel.band(...))
gd(basecol = 'tomato4')
p + layer( gpanel.band(..., col = 'grey10')) + layer(gpanel.fit(...))
p + layer( gpanel.band(...)) + layer(gpanel.fit(...))
# With panels and groups
z <- Prestige
z$gender <- with(z, cut( women, c(-1,15,50,101),labels = c("Male","Mixed","Female")))
tab(z, ~ gender + type)
z <- fillin( z, ~ education + type + gender, xpd = 1.1)
fit <- lm( income ~ (education+I(education^2)+I(education^3) )* type * gender,
z, na.action = na.exclude) # overfitting!
summary(fit)
z <- cbind( z, predict(fit, newdata = z, se = TRUE))
head(z)
(p <- xyplot( income ~ education| gender, z, groups = type,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
layout = c(1,3),
auto.key = list(space='right', lines= T, cex = 1.5)))
p + glayer( gpanel.band(...)) + glayer(gpanel.fit(...))
# trellis.focus()
# panel.identify(labels= z$occupation)
# trellis.unfocus()
z$type2 <- with( z, reorder(type,education, mean, na.rm=T))
gd(3)
(p <- xyplot( income ~ education| type2, z, groups = gender,
subscripts = T,
fit = z$fit,
lower = z$fit - z$se,
upper = z$fit + z$se,
layout = c(1,3),
par.strip.text = list(cex = 2),
auto.key = list(space='right', lines= T, cex = 1.5)))
p + glayer( gpanel.fit(...))
p + glayer( gpanel.band(...))
p + glayer( gpanel.band(...)) + glayer(gpanel.fit(...))
# trellis.focus()
# panel.identify(labels= z$occupation)
# trellis.unfocus()
# With panels^2
# need to remove 'col = col.line'
z <- Prestige
z$occ <- rownames(Prestige)
z$gender <- with(z, cut( women, c(-1,15,50,101),labels = c("Male","Mixed","Female")))
z$type2 <- with( z, reorder(type,education, mean, na.rm=T))
tab(z, ~ gender + type2)
z <- fillin( z, ~ education + type + gender, xpd = 1.1)
fit <- lm( income ~ (education+I(education^2)+I(education^3) )* type * gender,
z, na.action = na.exclude) # overfitting!
summary(fit)
z <- cbind( z, predict(fit, newdata = z, se = TRUE))
head(z)
(p <- xyplot( income ~ education| gender*type, z,
subscripts = T,
fit = z$fit,
labels = z$occ,
lower = z$fit - z$se,
upper = z$fit + z$se,
auto.key = list(space='right', lines= T, cex = 1.5)))
p + layer( gpanel.fit(...)) + layer( gpanel.band(...))
p + layer( gpanel.band(..., col = 'black', alpha = .1)) + layer(gpanel.fit(...)) +
layer(gpanel.text(...))
|
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.