Tagged Questions

A discrete, univariate distribution modelling the number of ${\rm Bernoulli}(p)$ trial successes until a specified number of failures occur.

learn more… | top users | synonyms

1
vote
1answer
12 views

Why does the null deviance in glm.nb differ between models of the same response variable?

I'm struggeling to understand the topic of deviance. Let's have two models as follows: ...
0
votes
0answers
13 views

Negative binomial instead of ordinal logistic?

I have data with one metric and a discrete outcome that I'm trying to estimate the probability of. The outcome is a count, but it's a safe assumption that the increments aren't i.i.d. I thought an ...
1
vote
0answers
26 views

“weight” input in glm.nb function in R. How exactly does the weight affect the likelihood?

I would like to understand how the weight argument of glm.nb is affecting the likelihood function. I understand that glm.nb find the MLE in an alternating iteration process where for a given theta the ...
2
votes
0answers
21 views

Is there a simple model that assumes a negative relationship between the mean and variance of response?

Poisson and Negative Binomial models assume that the variance of response is greater or equal to the mean of response. Is there a simple model where that assumption is reversed, i.e. variance goes ...
0
votes
0answers
12 views

A/B testing zero inflated data

I am trying to conduct an A/B(/C) test to compare the performance of 3 different website pages but I'm facing issues regarding zero inflated data. I have data for each page regarding ...
8
votes
1answer
125 views

Low sample size: LR vs F - test

Some of you might have read this nice paper: O’Hara RB, Kotze DJ (2010) Do not log-transform count data. Methods in Ecology and Evolution 1:118–122. klick. Currently I am comparing negative binomial ...
3
votes
2answers
42 views

Poisson as a limiting case of negative binomial

I was reading "Maximum Likelihood Estimation for the Negative Binomial Dispersion Parameter" by Walter W. Pieogorsch, and in the intro it says the Poisson distribution is a limiting case of negative ...
1
vote
0answers
51 views

Count data in a Structural Equation Model

I am currently trying to fit a structural equation model (SEM) on a certain dataset using R (lavaan package). Some of the most important variables in my model are count data (abundance of different ...
0
votes
1answer
12 views

Why addition of discrete probability values for negative binomial does not match to the cumulative probability value, when it matches in Poisson?

When I manually calculate just negative binomial probability it matches to the value from Excel, or R. formula : (nCr) * (p^k) * ( q^(x-k) ) x = 1000 ; k = 4 ; p = 0.01 ; [ and n = x-1 ; r = ...
0
votes
0answers
20 views

How to interpret logarithmically transformed coefficients in negative binomial regression?

How can I interpret log-transformed independent variables in terms of percent change in a negative binomial regression?
5
votes
0answers
41 views

Is the negative binomial not expressible as in the exponential family if there are 2 unknowns?

I had a homework assignment to express the negative binomial distribution as an exponential family of distributions given that the dispersion parameter was a known constant. This was fairly easy, but ...
2
votes
1answer
49 views

When to use zero-inflated poisson regression and negative binomial distribution

I have a fairly simple dataset looking at the relationship between the first nesting date of a bird in a given year (Date) and the birds overall fledgling production from that year (Fledge; count data ...
3
votes
0answers
30 views

Interpretation of log transformed predictor in negative binomial regression

I mainly want to make sure that I'm making the correct interpretation here. I built a negative binomial regression model predicting a count variable. There was evidence of overdispersion or I would ...
2
votes
1answer
62 views

Quantiles of a compound gamma/negative binomial distribution

Are there any formulae for the quantiles of a compound gamma/negative binomial distribution? That is, suppose we have $$N \sim \text{NegBin}(\alpha, \lambda)$$ and conditional on $N$, $$Y = ...
0
votes
1answer
70 views

Interpretation of $\theta$ in negative binomial regression

First off, a very similar question has been asked before. But the answers to this question did not explain what high/low values of theta mean. Here's my crack at trying to figure out what high/low ...
1
vote
1answer
29 views

Zero-inflated negative binomial model for true zeros

The zeroinfl function in the pscl package in R assumes that zeros include both false zeros and true zeros. I have a zero ...
9
votes
1answer
126 views

Increased Type I error - GLM

Some of you might have read this nice paper: O’Hara RB, Kotze DJ (2010) Do not log-transform count data. Methods in Ecology and Evolution 1:118–122. klick. In my field of research (ecotoxicology) ...
4
votes
1answer
57 views

Generate pseudo-random overdispersed Poisson numbers

I have multiple sets of data which conform to overdispersed Poisson distributions which I can model with the alternative parameterization of a negative binomial distribution ($\mu$ and $D$ instead of ...
2
votes
1answer
72 views

GLM regression - help choosing model specification

I think I need to use a Poisson-family regression or negative binomial regression. My variables are as follows: Y is an integer value ranging from 0 to ~1200. It represents sums (number of species ...
0
votes
1answer
49 views

Validating residual plot count data (different levels)

I am studying the distribution of a marine species using the number of sightings as a dependent variable. When I am trying to validate the plots of the best model I am getting a non-usual pattern, and ...
3
votes
1answer
47 views

Use law of total variance to find unconditional variance of overdispersed Poisson?

First, I need to prove that the distribution of a RV X, where X|lambda ~ Pois(lambda), and lambda ~ gamma(a, B), is a negative binomial. I know that it is, but why negative binomial instead of another ...
0
votes
1answer
20 views

Does and 95% CI containing 1 imply no difference in IRR calculations in the same was as in OR calculations?

My understanding is that with an odds ratio (OR) that a 95% CI containing 1 means that there is no difference between the odds of something occurring between groups. I wondered if this holds true for ...
0
votes
0answers
26 views

Bootstrap of parameter estimates and confidence intervals (hurdle model)

Think you can help with this. I´ve run a set of candidate hurdle models on insect abundance data with pscl package for R. These models had an abundance part with a truncated negative binomial ...
2
votes
2answers
107 views

How to evaluate goodness of fit for negative binomial regression

I'm trying to fit a model estimating waiting time using negative binomial regression, but I'm not sure how to assess the goodness of fit for my model. I would like to compare the negative binomial ...
0
votes
1answer
50 views

correct use of Negative Binomial with a Geometric distribution in a mixed model (glmmPQL)

I am trying to fit a NB GLMM with a gemoetric distribution. I have come across very little information on this form of regression. And would like some pointers/reasurance. some literature is ...
2
votes
0answers
51 views

does the equivalence of Poisson and conditional logit models hold for fixed effect panel data and negative binomial models?

Guimaraes et al. (Rev Econ Stat, 2003, 85/1) describe the conditions under which the results from poisson regression models and conditional logit models are equivalent. I am trying to find out whether ...
2
votes
0answers
60 views

Basic idea of zero inflated two part models(hurdel) and application to big data (machine learning)

I'm currently working on the data which has 90% 0s in response variable. Based on my research, it seems zero inflated models could be a solution to this. However, while I was reading related ...
0
votes
0answers
105 views

R² (squared) from a generalized linear mixed-effects models (GLMM) using a negative binomial distribution

I try to compute the marginal and conditional R² for a GLMM using a negative binomial distribution by following the procedure recommended by Nakagawa & Schielzeth (2013) . Unfortunately, the ...
2
votes
0answers
141 views

Estimating abundance using non-normal count data

I have sample counts of $n=20$ or $n=7$ taken from right-skewed and zero-inflated populations. The challenge in each case is to use the sample to estimate the total count in that population. Each of ...
1
vote
0answers
23 views

Limits of zero-inflated negative binomial in % of zeroes

What are the limits of a zero-inflated regression? Specifically, if more than 80% of the data is zeroes, is the ZINB still valid? What is a good rule of thumb or educated way to understand the right ...
5
votes
1answer
82 views

Zero inflated negative binomial with selection

I am looking for a Stata (or R/Matlab if there's no Stata) implementation of the model described by Greene (1994) (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1293115). It is essentially a ...
0
votes
0answers
11 views

comparing same negative binomial models built with subset data

I have a dataset made our of several stacked datasets (one for each state). I want to check whether a zero inflated negative binomial model with data from an individual state is different from the ...
1
vote
0answers
68 views

GAM with negative binomial and temporal autocorrelation

I want to fit a GAM in R. My data are negative binomial. Plus they have a temporal autocorrelation. ...
0
votes
1answer
32 views

Controlling for variables in negative binomial regression

I would like to run a negative binomial regression, but also controlling for one of the variables. Similar to what you would do with a hierarchical regression. I am trying to predict violence (never, ...
1
vote
0answers
24 views

Leave-one-out cross-validation for negative binomial

I've just started working with the GLM negative binomial and I was wondering if it's possible to carry out a leave-one-out cross-validation for negative binomial GLM as I'm having troubles doing it ...
6
votes
1answer
94 views

In a Poisson process measured with some efficiency, is the measured count still Poisson?

Situation: Say I have a Poisson process, like radioactive decay, producing R particles per second. I measure with a detector. There is a probability P that a particle will be detected by the ...
2
votes
1answer
21 views

I got a good OLS fit for integer variables, do I still need to use count data methods?

First of all, I'm not a statistician, but I'm teaching myself some methods I require for a project I'm doing now. I have a 2D dataset of N observations. For the ith observation, the first entry is ...
1
vote
0answers
45 views

Can i use a mixture model for when I have an omitted variable?

I plan to fit a GAM or GAMM. There is one categorical variable which I think is important for explaining Y (or Y*), but it is not in my dataset - it is measurable but has not been measured. Can I use ...
8
votes
2answers
392 views

Why are Pearson's residuals from a negative binomial regression smaller than those from a poisson regression?

I have these data: set.seed(1) predictor <- rnorm(20) set.seed(1) counts <- c(sample(1:1000, 20)) df <- data.frame(counts, predictor) I ran a poisson ...
0
votes
0answers
78 views

fisher information matrix of Negative Binomial distribution

can anyone show me how to find the Fisher information matrix for negative binomial distribution, if I parameterize the Negative Binomial distribution using parameters mean and size. Thank you!
2
votes
2answers
285 views

Why do ANOVAs and GLM (negative binomial model) give different results for interaction effects?

In my study, children repeat sentences and I count the errors (DV = error rate). Children are divided into two groups (Group factor), and there are two different types of sentences (Condition factor). ...
3
votes
1answer
226 views

Zero-inflation on steroids: choose among Poisson, negative binomial and zero-inflated regressions

I am struggling to fit alternative count models into my data. I guess my problem is just too many zeros. This is my data ...
1
vote
1answer
100 views

Negative binomial distribution fit

I am trying to fit a negative binomial distribution, in R, to my over dispersed data (out of 20 ,14 samples are 0, and rest are less than 5). The mean is $-0.8$ and ...
2
votes
0answers
225 views

Hyper-prior for negative binomial in hierarchical model using JAGS/BUGS

Below I'm using a negative binomial because it is more flexible than a simple poisson model. The data are counts $y$ of events for 16 individuals $x$. There are 14 counts (i.e. counting periods) for ...
3
votes
2answers
80 views

Does the estimated overdispersion parameter of Negative Binomial depend on mean

Negative Binomial distribution can be parameterized using mean, $\mu$, and overdispersion $\psi$, so that the variance of NB is $\mu + \frac{\mu^2}{\psi}$. We know there is no analytical solution for ...
3
votes
0answers
49 views

Evaluate deviation from negative binomial model

I'm trying to figure out how to determine to what extent a sample deviates from a negative binomial model fitted to a larger population. As an example, I generated counts of doctor visits for a ...
2
votes
0answers
180 views

Standardizing count variables in panel data with overdispersion - R or Stata

I'm running a regression where the dependent (response) variable is a highly dispersed (slightly zero-inflated) count and the explanatory (independent or predictor) variables are continuous, counts as ...
3
votes
0answers
88 views

Error in fitting negative binomial regression model in R when replicating published results (works in Stata)

I'm trying to replicate the results of the first model of this article: Hultman, Lisa, Jacob Kathman, and Megan Shannon. 2013. “United Nations Peacekeeping and Civilian Protection in Civil War.” ...
0
votes
1answer
39 views

Interpreting ZINB - inflation model non-significant

I have a zero-inflated negative binomial model to a dataset (n = 47) with a over-dispersed dependent variable (...
1
vote
1answer
97 views

Parametrization of Gamma and Negative Binomial in R

I have some Poisson data {${y_1,...,y_n}$} and a Gamma prior, and I wish to construct a predictive posterior distribution. As I understand, if my Gamma hyperparameters are $\alpha$ (the prior number ...