[英]Overlay of forest plot from ZINB model
我想要使用sjPlot
包覆盖完整 ZINB 模型和数据子集的森林图。 如您所知,ZINB 模型产生两种模型:一种用于计数模型,一种用于零膨胀模型。 当从完整数据或数据子集使用 ZINB 模型时, plot_model
工作正常,这意味着为两个模型(计数和零模型)生成一个图,但是当我使用plot_models叠加时,只为计数模型生成一个图。 我正在寻找完整模型和子模型中的计数和零膨胀模型图,用于完整数据和数据子集。 任何帮助将非常感激
library(sjPlot)
library(sjlabelled)
library(sjmisc)
library(ggplot2)
library(MASS)
library(pscl)
library(boot)
zinb_all_uni <- zeroinfl(ivdays~age,
link="logit",
dist = "negbin",
data=caterpillor)
summary(zinb_all_uni)
plot_model(zinb_all_uni, type="est")
zinb_full_adj <- zeroinfl(ivdays~age+sex+edu,
link="logit",
dist = "negbin",
data=caterpillor)
summary(zinb_full_adj)
plot_model(zinb_full_adj, type="est", terms = c("count_ageb", "count_agec", "zero_ageb", "zero_agec"))
############ second model#######
Zinb_uni_sub <- zeroinfl(ivdays~age,
link="logit",
dist = "negbin",
data=subset(caterpillor, country=="eng"))
summary(zinb_uni_sub)
plot_model(zinb_uni_sub, type="est")
zinb_adj_sub <- zeroinfl(ivdays~age+sex+edu,
link="logit",
dist = "negbin",
data=subset(caterpillor, country=="eng"))
summary(zinb_adj_sub)
plot_model(zinb_adj_sub, type="est", terms = c("count_ageb", "count_agec", "zero_ageb", "zero_agec"))
### overlying plots from both models
plot_models(zinb_all_uni, Zinb_uni_sub)
plot_models(zinb_full_adj, zinb_adj_sub)
数据:
caterpillor=structure(list(id = 1:100,
age = structure(c(1L, 1L, 2L, 1L,
2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 3L, 3L, 3L,
1L, 1L, 2L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 2L, 1L, 2L, 2L,
2L, 3L, 3L, 3L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L,
2L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 3L,
3L, 3L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 2L, 1L,
2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 3L, 3L, 3L),
.Label = c("a", "b", "c"), class = "factor"),
sex = structure(c(2L,
1L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 1L, 1L, 1L,
1L, 2L, 1L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 1L,
1L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L, 2L,
2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 1L,
2L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L,
2L, 2L, 1L, 1L, 1L, 1L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 2L, 2L,
2L, 1L, 1L),
.Label = c("F", "M"), class = "factor"),
country = structure(c(1L,
1L, 1L, 1L, 3L, 3L, 3L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 3L, 3L, 3L,
2L, 2L, 2L, 1L, 1L, 1L, 1L, 3L, 3L, 3L, 2L, 2L, 2L, 1L, 1L, 1L,
1L, 3L, 3L, 3L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 3L, 3L, 3L, 2L, 2L,
2L, 1L, 1L, 1L, 1L, 3L, 3L, 3L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 3L,
3L, 3L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 3L, 3L, 3L, 2L, 2L, 2L, 1L,
1L, 1L, 1L, 3L, 3L, 3L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 3L, 3L, 3L,
2L, 2L, 2L),
.Label = c("eng", "scot", "wale"), class = "factor"),
edu = structure(c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L,
1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L,
1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L,
1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L),
.Label = c("x", "y", "z"), class = "factor"),
lungfunction = c(45L,
23L, 25L, 45L, 70L, 69L, 90L, 50L, 62L, 45L, 23L, 25L, 45L,
70L, 69L, 90L, 50L, 62L, 45L, 23L, 25L, 45L, 70L, 69L, 90L,
50L, 62L, 45L, 23L, 25L, 45L, 70L, 69L, 90L, 50L, 62L, 45L,
23L, 25L, 45L, 70L, 69L, 90L, 50L, 62L, 45L, 23L, 25L, 45L,
70L, 69L, 90L, 50L, 62L, 45L, 23L, 25L, 45L, 70L, 69L, 90L,
50L, 62L, 45L, 23L, 25L, 45L, 70L, 69L, 90L, 50L, 62L, 45L,
23L, 25L, 45L, 70L, 69L, 90L, 50L, 62L, 25L, 45L, 70L, 69L,
90L, 50L, 62L, 25L, 45L, 70L, 69L, 90L, 50L, 62L, 25L, 45L,
70L, 69L, 90L),
ivdays = c(15L, 26L, 36L, 34L, 2L, 4L, 5L,
8L, 9L, 15L, 26L, 36L, 34L, 2L, 4L, 5L, 8L, 9L, 15L, 26L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 5L, 8L, 9L, 36L, 34L, 2L, 4L, 5L, 8L,
9L, 36L, 34L, 2L, 4L, 5L),
no2_quintile = structure(c(1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L,
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L,
5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L),
.Label = c("q1", "q2",
"q3", "q4", "q5"), class = "factor")),
class = "data.frame", row.names = c(NA,
-100L))
但是当我叠加地块时,我只得到一个地块
代码如下,基本要点:
plot_model
这样的自动化机器时遇到麻烦时,我通常更喜欢使用像broom::tidy()
(用于系数)或ggeffects
或emmeans
包(用于预测)这样的机器并构建我自己的 ggplot — 对我来说,这比尝试更容易弄清楚自动化程度更高的工具在做什么broom
没有用于zeroinfl
模型的tidy()
方法,但通过谷歌搜索可以在poissonreg
包中找到一个方法...tidy()
方法没有用于构建置信区间或将系数反向转换为计数比或优势比标度的机制,因此我必须在下面实现自己的...library(broom)
library(poissonreg)
library(tidyverse) ## purrr::map_dfr, ggplot ...
theme_set(theme_bw())
library(colorspace)
mod_list <- list(all_uni = zinb_all_uni, uni_sub = Zinb_uni_sub,
full_adj = zinb_full_adj, adj_sub = zinb_adj_sub)
tidy(zinb_all_uni, type = "all")
coefs <- (mod_list
|> map_dfr(tidy, type = "all",
.id = "model")
## construct CIs
|> mutate(conf.low = qnorm(0.025, estimate, std.error),
conf.high = qnorm(0.975, estimate, std.error))
|> filter(term != "(Intercept)") ## usually don't want this
## cosmetic (strip results down to the components we actually need)
|> select(model, term, type, estimate, conf.low, conf.high)
## back-transform
|> mutate(across(c(estimate, conf.low, conf.high), exp))
)
ggplot(coefs, aes(x = estimate, y = term, colour = model)) +
geom_pointrange(aes(xmin = conf.low, xmax = conf.high),
position = position_dodge(width = 0.5)) +
## separate count-ratio and odds-ratio (conditional/zero) plots
facet_wrap(~type, scale = "free") +
scale_color_discrete_qualitative() ## cosmetic
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.