r - R 中的负二项式：glm.nb：在 sqrt(1/i) 中：产生了 NaN，以及其他问题

Question

我正在运行负二项式回归。我想知道为什么会出现以下错误：

在 sqrt(1/i) 中：产生的 NaN

似乎“i”中有一些负值，但我该如何避免呢？

另一个是：

在 loglik(n, th, mu, Y, w) 中：“lgamma”中的值超出范围

这可能是第一个错误的后果，所以如果我修复第一个错误，第二个可能会消失。或者可能不是。

在其他一些情况下，我能够计算回归，但以下输出对我来说似乎很奇怪：

(Dispersion parameter for Negative Binomial(10684331573) family taken
to be 1)

    Null deviance: 8779.49  on 359  degrees of freedom 
Residual deviance:  270.32  on 200  degrees of freedom
AIC: 2074.7

Number of Fisher Scoring iterations: 1


              Theta:  10684331573
          Std. Err.:  615849693813

 2 x log-likelihood:  -1752.749

这些数字看起来好吗？我的意思是色散参数、theta 和标准误差。它们对我来说看起来非常大，因此我不确定结果是否还可以。

使用泊松回归时我从来没有遇到过类似的问题，但后来我意识到我的数据过于分散，这就是我使用负二项式的原因。但是，我在这方面遇到了很多麻烦。

这是代码：

negbin <- glm.nb(Freq ~ cluster*gender*agecombined*educ, maxit=100)
mod.good <- step(negbin, direction='both', maxit=100)

这是dput整个数据集的：

structure(list(gender = structure(c(1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,
2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L,
1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,
2L, 1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L,
1L,  2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L,
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L, 1L, 2L, 1L, 2L,
1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L,  2L), .Label = c("1",
"2"), class = "factor"), agecombined = structure(c(1L,  1L, 2L, 2L,
3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L,  3L, 4L, 4L, 5L,
5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L,  5L, 6L, 6L, 1L, 1L,
2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L,  1L, 2L, 2L, 3L, 3L, 4L,
4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L,  3L, 4L, 4L, 5L, 5L, 6L, 6L,
1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L,  5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L,
3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L,  1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L,
6L, 6L, 1L, 1L, 2L, 2L, 3L,  3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L,
2L, 3L, 3L, 4L, 4L, 5L,  5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L,
5L, 5L, 6L, 6L, 1L,  1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L,
1L, 2L, 2L, 3L,  3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L,
4L, 4L, 5L,  5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L,
6L, 1L,  1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L,
3L,  3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 
5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L,  1L,
2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L,  3L, 4L,
4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L,  5L, 6L, 6L,
1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L,  1L, 2L, 2L, 3L,
3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L,  3L, 4L, 4L, 5L, 5L,
6L, 6L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L,  5L, 6L, 6L, 1L, 1L, 2L,
2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L, 6L, 1L,  1L, 2L, 2L, 3L, 3L, 4L, 4L,
5L, 5L, 6L, 6L, 1L, 1L, 2L, 2L, 3L,  3L, 4L, 4L, 5L, 5L, 6L, 6L),
.Label = c("18-24", "25-34", "35-44",  "45-54", "55-64", "65 and
older"), class = "factor"), educ = structure(c(1L,  1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L,  2L, 2L, 2L, 2L, 2L,
2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,  3L, 3L, 3L, 4L, 4L, 4L,
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L,  5L, 5L, 5L, 5L, 5L, 5L, 5L,
5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L,  6L, 6L, 6L, 6L, 6L, 6L, 6L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,  1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 3L,  3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,
3L, 4L, 4L, 4L, 4L, 4L,  4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L,
5L, 5L, 5L, 5L, 5L,  5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L,
6L, 6L, 6L, 1L,  1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L,
2L, 2L, 2L,  2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,
3L, 3L,  3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L,
5L,  5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L, 
6L, 6L, 6L, 6L, 6L, 6L, 6L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,  1L,
1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L,  3L, 3L,
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L,  4L, 4L, 4L,
4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L,  5L, 5L, 5L, 6L,
6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 1L,  1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L,  2L, 2L, 2L, 2L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,  3L, 3L, 3L, 4L, 4L, 4L, 4L,
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L,  5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L,
5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L,  6L, 6L, 6L, 6L, 6L, 6L, 6L), .Label =
c("2-year college", "BA",  "Illiterate", "MA or higher", "Primary",
"Secondary"), class = "factor"), 
    cluster = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 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, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
    2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
    2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 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, 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, 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, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 4L, 4L, 4L, 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, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
    5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
    5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
    5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 
    5L, 5L, 5L, 5L, 5L, 5L), .Label = c("E", "A", "B", "C", "D"
    ), class = "factor"), Freq = c(27L, 18L, 48L, 29L, 18L, 19L, 
    14L, 10L, 2L, 1L, 2L, 0L, 48L, 36L, 69L, 54L, 33L, 15L, 12L, 
    4L, 5L, 1L, 0L, 0L, 2L, 4L, 12L, 14L, 17L, 17L, 23L, 32L, 
    16L, 17L, 18L, 6L, 4L, 2L, 17L, 7L, 8L, 4L, 5L, 0L, 1L, 0L, 
    0L, 0L, 53L, 42L, 82L, 58L, 81L, 60L, 42L, 35L, 16L, 14L, 
    22L, 6L, 83L, 40L, 62L, 54L, 43L, 46L, 26L, 12L, 15L, 3L, 
    3L, 3L, 11L, 13L, 11L, 23L, 16L, 18L, 11L, 5L, 1L, 3L, 1L, 
    1L, 26L, 44L, 34L, 54L, 25L, 41L, 19L, 17L, 10L, 3L, 3L, 
    0L, 4L, 4L, 7L, 14L, 22L, 31L, 14L, 34L, 14L, 33L, 14L, 20L, 
    7L, 11L, 22L, 11L, 14L, 8L, 8L, 1L, 2L, 0L, 1L, 2L, 29L, 
    65L, 34L, 84L, 36L, 65L, 28L, 39L, 16L, 15L, 16L, 9L, 25L, 
    51L, 12L, 38L, 23L, 29L, 22L, 19L, 7L, 5L, 5L, 1L, 7L, 16L, 
    14L, 35L, 6L, 27L, 8L, 5L, 1L, 1L, 1L, 0L, 24L, 57L, 29L, 
    53L, 24L, 28L, 11L, 9L, 7L, 2L, 0L, 0L, 3L, 7L, 1L, 8L, 2L, 
    18L, 5L, 13L, 10L, 11L, 5L, 10L, 3L, 1L, 5L, 13L, 4L, 2L, 
    2L, 1L, 1L, 0L, 0L, 0L, 14L, 51L, 21L, 77L, 23L, 50L, 25L, 
    31L, 17L, 16L, 13L, 13L, 19L, 52L, 24L, 59L, 18L, 44L, 9L, 
    20L, 6L, 3L, 7L, 2L, 14L, 28L, 34L, 47L, 29L, 47L, 15L, 13L, 
    9L, 3L, 2L, 0L, 46L, 75L, 124L, 81L, 67L, 45L, 33L, 15L, 
    9L, 4L, 5L, 3L, 0L, 10L, 6L, 19L, 12L, 28L, 22L, 37L, 31L, 
    41L, 26L, 31L, 7L, 6L, 21L, 13L, 6L, 7L, 8L, 2L, 2L, 1L, 
    0L, 0L, 67L, 89L, 116L, 159L, 99L, 102L, 64L, 80L, 42L, 25L, 
    25L, 8L, 108L, 123L, 60L, 97L, 68L, 66L, 44L, 35L, 12L, 5L, 
    9L, 2L, 7L, 3L, 53L, 15L, 33L, 3L, 8L, 3L, 4L, 0L, 0L, 0L, 
    48L, 19L, 76L, 40L, 55L, 11L, 16L, 1L, 4L, 1L, 2L, 0L, 6L, 
    7L, 21L, 22L, 18L, 23L, 32L, 37L, 40L, 13L, 23L, 10L, 4L, 
    2L, 19L, 2L, 8L, 3L, 6L, 0L, 1L, 0L, 1L, 0L, 68L, 37L, 90L, 
    42L, 76L, 38L, 47L, 16L, 29L, 5L, 18L, 2L, 82L, 32L, 62L, 
    27L, 44L, 22L, 20L, 8L, 8L, 2L, 1L, 0L)), .Names = c("gender",  "agecombined", "educ", "cluster", "Freq"), row.names = c(NA, 
-360L), class = "data.frame")

r - R 中的负二项式：glm.nb：在 sqrt(1/i) 中：产生了 NaN，以及其他问题

0 回答 0

Related

Reference