并排打印 glmmTMB 条件模型和零膨胀模型与模型摘要

Question

我正在尝试打印使用 glmmTMB 和 modelsummary 估计的零膨胀泊松模型的条件和零膨胀模型，但我无法这样做。

这是我迄今为止的尝试，遵循here

描述的语法

Owls <- transform(Owls,
                  Nest=reorder(Nest,NegPerChick),
                  NCalls=SiblingNegotiation,
                  FT=FoodTreatment)

fit_zipoisson <- glmmTMB(NCalls~(FT+ArrivalTime)*SexParent+
                           offset(log(BroodSize))+(1|Nest),
                         data=Owls,
                         ziformula=~1,
                         family=poisson)

ti <- list(
  broom.mixed::tidy(fit_zipoisson) |>
    filter(effect=="fixed" & component=="cond"),
  broom.mixed::tidy(fit_zipoisson) |>
    filter(effect=="fixed" & component=="zi")
)

gl <- list(
  broom.mixed::glance(fit_zipoisson),
  broom.mixed::glance(fit_zipoisson)
)

mod <- list(
  tidy = ti,
  glance = gl)
class(mod) <- "modelsummary_list"

modelsummary(mod)

返回：

Error: `estimate` is not available. The `estimate` and `statistic` arguments must correspond to column names in the output of this command: `get_estimates(model)`

我猜这个语法不支持传递列表？

Answer 1

您可以使用

shape

参数使用

get_estimates()

函数返回的列名来构造表。

library(modelsummary)
library(glmmTMB)

Owls <- transform(Owls,
  Nest = reorder(Nest, NegPerChick),
  NCalls = SiblingNegotiation,
  FT = FoodTreatment)

fit <- glmmTMB(
  NCalls ~ (FT + ArrivalTime) * SexParent +
    offset(log(BroodSize)) + (1 | Nest),
  data = Owls,
  ziformula = ~1,
  family = poisson)

get_estimates(fit)
#>                        term    estimate  std.error conf.level
#> 1               (Intercept)  2.53994692 0.35656284       0.95
#> 2                FTSatiated -0.29110639 0.05960977       0.95
#> 3               ArrivalTime -0.06807809 0.01427062       0.95
#> 4             SexParentMale  0.44884508 0.45002291       0.95
#> 5  FTSatiated:SexParentMale  0.10472505 0.07286248       0.95
#> 6 ArrivalTime:SexParentMale -0.02139750 0.01834893       0.95
#> 7               (Intercept) -1.05753356 0.09411867       0.95
#> 8       SD (Intercept Nest)  0.35966420         NA       0.95
#>      conf.low   conf.high   statistic df.error      p.value effect
#> 1  1.84109659  3.23879725   7.1234201      Inf 1.052813e-12  fixed
#> 2 -0.40793938 -0.17427339  -4.8835350      Inf 1.042007e-06  fixed
#> 3 -0.09604799 -0.04010818  -4.7705056      Inf 1.837640e-06  fixed
#> 4 -0.43318360  1.33087377   0.9973827      Inf 3.185788e-01  fixed
#> 5 -0.03808278  0.24753288   1.4372974      Inf 1.506335e-01  fixed
#> 6 -0.05736075  0.01456575  -1.1661442      Inf 2.435562e-01  fixed
#> 7 -1.24200277 -0.87306435 -11.2361715      Inf 2.708767e-29  fixed
#> 8  0.25483237  0.50762131          NA       NA           NA random

因此我们使用

component

列来对参数进行排序。请注意，公式的顺序对于表的结构很重要，正如手册页和

中详细记录的那样

modelsummary(list(fit, fit),
  shape = component + statistic ~ model)

+-----------------------------+---------------+---------+---------+
|                             | component     | (1)     | (2)     |
+=============================+===============+=========+=========+
| (Intercept)                 | conditional   | 2.540   | 2.540   |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.357) | (0.357) |
+-----------------------------+---------------+---------+---------+
|                             | zero_inflated | -1.058  | -1.058  |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.094) | (0.094) |
+-----------------------------+---------------+---------+---------+
| FTSatiated                  | conditional   | -0.291  | -0.291  |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.060) | (0.060) |
+-----------------------------+---------------+---------+---------+
| ArrivalTime                 |               | -0.068  | -0.068  |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.014) | (0.014) |
+-----------------------------+---------------+---------+---------+
| SexParentMale               |               | 0.449   | 0.449   |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.450) | (0.450) |
+-----------------------------+---------------+---------+---------+
| FTSatiated × SexParentMale  |               | 0.105   | 0.105   |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.073) | (0.073) |
+-----------------------------+---------------+---------+---------+
| ArrivalTime × SexParentMale |               | -0.021  | -0.021  |
+-----------------------------+---------------+---------+---------+
|                             |               | (0.018) | (0.018) |
+-----------------------------+---------------+---------+---------+
| SD (Intercept Nest)         |               | 0.360   | 0.360   |
+-----------------------------+---------------+---------+---------+
| Num.Obs.                    |               | 599     | 599     |
+-----------------------------+---------------+---------+---------+
| R2 Marg.                    |               | 0.015   | 0.015   |
+-----------------------------+---------------+---------+---------+
| R2 Cond.                    |               | 0.066   | 0.066   |
+-----------------------------+---------------+---------+---------+
| AIC                         |               | 4015.6  | 4015.6  |
+-----------------------------+---------------+---------+---------+
| BIC                         |               | 4050.8  | 4050.8  |
+-----------------------------+---------------+---------+---------+
| ICC                         |               | 0.1     | 0.1     |
+-----------------------------+---------------+---------+---------+
| RMSE                        |               | 5.96    | 5.96    |
+-----------------------------+---------------+---------+---------+