Questions tagged [generalized-linear-model]
A generalization of linear regression allowing for nonlinear relationships via a "link function" and for the variance of the response to depend on the predicted value. (Not to be confused with "general linear model" which extends the ordinary linear model to general covariance structure and multivariate response.)
4,503
questions
3
votes
1
answer
38
views
What to do when the residuals of a general LMM are non normal and a generlized LMM will not build, when is it just too non-normal?
I've been having no luck building a general LMM. The residuals are not normally distributed, they follow somewhat of a leptokurtic distribution and homoskedacity is also present. Using log ...
- 83
0
votes
0
answers
21
views
Signs for marginal effects and regression coefficients are inconsistent for beta regression
I have conducted a zero-one inflated beta (ZOIB) regression using a logit link function for explaining tenure incidence in colleges and universities. Tenure incidence is a proportion in the interval [...
0
votes
0
answers
51
views
Why does weights argument in glm change the outcome of lsd?
I have dataframe df.
...
0
votes
0
answers
5
views
link log and identity in GLMER
Say that we have a GLM model with the following formula:
outcome = b1x1 + b2x2 + b0
and outcome is cost, x1, x2 are independent variables
Fitted using log link with gaussian distribution, so log ...
0
votes
1
answer
25
views
GLM vs Kruskal-Wallis/ANOVA?
I am trying to analyse some data in R that has one continuous dependent variable (x) and two categorical variables (sex: M/F, surface: D/V). My main goal is to understand if sex and surface affect 'x'....
- 11
0
votes
0
answers
29
views
GLM or GLMM design for different tasks
I am conducting a study involving 70 participants diagnosed with Mild Cognitive Impairment (MCI) and 75 without MCI. All participants were engaged in taxonomic semantic, thematic semantic, and ...
2
votes
1
answer
55
views
Distribution of a conical combination of n poisson variables?
Does a conical combination of n Poisson distributed variables have a closed-form distribution (linear combination with nonnegative coefficients)? I know that the sum of random Poisson variables would ...
- 3,080
1
vote
0
answers
48
views
Issue of multicollinearity in R for glm analysis
I was wondering if someone could help me with a statistical problem I have run into. Any help would be incredibly helpful.
Please note that for clarity, I have simplified the below description. It ...
- 63
2
votes
0
answers
87
views
How can we justfify the assumption of equal scale/variance in the definition of R-squared from Deviances in GLMs?
If we take the R-squared to be the comparison of Deviances between models (the model of interest, the saturated model, and the constant model), we can write it as (see this answer CC BY-SA 4.0):
$$R_{...
- 18.4k
1
vote
0
answers
32
views
Model test sensitivity to ability
I have a set of tests and a population of agents whose ability I want to assess. Each agent has taken some of the tests. The agents have no memory of the tests they have taken, so each time an agent ...
- 173
2
votes
1
answer
91
views
What type of regression to use when outcome is integers from 0 to 10
I have an outcome variable that measures "community perception," with responses ranging from $0$ to $10$. For instance, the outcome variable might represent answers to the question: "In ...
- 23
0
votes
0
answers
23
views
Drop Level for Categorical Variables in GLM [closed]
I am running a regression with some numerical and some categorical variables. This issue that I am encountering is that 2 levels of categorical variables are perfectly correlated.
For example, let's ...
4
votes
1
answer
79
views
How to structure repeated measures in GLM based on my study design? Nested or not nested
So I'm having some trouble deciding how to format the repeated measures in my model. Basically deciding between:
...
- 41
1
vote
1
answer
63
views
How can Null model likelihood be higher than Fitted model likelihood
As far as I know, when fitting a GLM, the fitted model should always have a higher likelihood compared to the null model (with only an intercept) for the same training set. When I run a small ...
- 1,323
2
votes
1
answer
42
views
Appropriate way to configure contrasts and interactions in GLM likelihood ratio test
I have a pretty large experiment where we are looking for differences in gene expression in RNA sequencing experiments. We are using EdgeR's GLM functionality, and then using contrasts to look for ...
- 121
2
votes
1
answer
66
views
Prove that the Deviance and the Generalised Pearson Statistic are asymptotically equivalent
I am reading the paper Exponential Dispersion Models from Jørgesen and at page $137$ I have encountered a claim that I don't know how to prove.
The author claims that the Generalised Pearson Statistic,...
- 192
0
votes
0
answers
25
views
Before-after intervention variation: difference in-difference or broken-stick regression?
I am trying to figure out which analysis would best answer our research question.
Our study aimed to find out whether the variation in surgical rate decreases in a group of hospitals after an ...
0
votes
0
answers
5
views
Simulating a dataset from model output when model includes multiple binary deviation-coded variables
I am trying to simulate data using parameters from a glmer() model output. The model, which comes from a published paper, is as follows: DV ~ 1 + group* sex *verb type + trial number + (1 |participant)...
2
votes
1
answer
41
views
Pseudo $R^2$ for probit model: In-sample or out-of-sample?
I have a dataset test_data that measures mortality in response to dosage of a pesticide. I used a probit model that evaluates the efficacy of a single pesticide. ...
0
votes
0
answers
20
views
How to handle explanatory variables with meaningful missing/NA values in logistic regression [duplicate]
I am wanting to fit a logistic regression where some explanatory variables have null values. The nulls are meaningful - for instance, a continuous explanatory variable capturing 'time since last ...
- 220
0
votes
0
answers
21
views
Why do the standardized beta values and CIs of a glm poisson regression model not differ from the unstandardized ones (using report function)?
For a specific research question i fitted a generalized linear mixed model using a poisson link function due to the characteristics of my data.
For reporting purposes i used the report package and the ...
1
vote
0
answers
38
views
GLM: invalid value encountered in log special.gammaln
I've never used GLM before so I would like to have some hints on how to use it and if I'm missing any steps.
My challenge:
I want to know if the price of product is influenced, positively or ...
- 111
0
votes
0
answers
24
views
Interpreting and transforming GLM output parameters with a Gamma log link
I built a GLM model in R with a Gamma log link and where my response variable is "1 - effectiveness". I would like to report the results of my model directly in terms of "effectiveness&...
- 123
0
votes
1
answer
65
views
Rule of thumb for deciding between Poisson and negative binomal models
I am analyzing the number of a specific class of mutations in cancer genome sequencing data, as a function of: (1) overall mutation rate per sample, (2) length of the genomic segment (gene) being ...
- 113
-1
votes
1
answer
59
views
Is there a link function $g$ which is a monotonic function with $g(0) = 0$? [closed]
I am working on a machine learning project (similar to this paper https://arxiv.org/abs/1912.04136). For this I need to make an assumption that a link function is monotonic and takes value $0$ at $0$. ...
- 281
1
vote
0
answers
18
views
Interpreting the results/output of a lme (linear mixed effects) model
I'm conducting an experiment where I'm trying to determine if a response variable (here, met rate) is affected by treatment (there are 3 treatment groups: CON, FLUX, and HOT). Here, the metrate was ...
- 11
0
votes
0
answers
15
views
glmmTMB profile argument, what does it do?
I tried running analysis using glmmTMB with profile=TRUE and the model that previously wasn't converging, now converging, anyone knows what assumptions does this parameter "profile" takes ...
- 71
0
votes
0
answers
23
views
GLM with nested predictor and non-nested outcome
I am using GLM to analyze the effect of career type (4-category discrete) on performance score in a competition (1-100 continuous). I am interested to know whether a particular career type results in ...
- 101
1
vote
1
answer
52
views
Replacing Hypothesis Testing with Regression
It appears, unless I am misunderstanding something fundamental, there is a way to preform hypothesis testing using a linear model instead. Is there any simple source (book?) that shows how to do this? ...
- 2,080
1
vote
2
answers
90
views
Do all GLM models not require equal variance?
I am trying to learn about generalized linear models (GLMs). For example, in a Poisson GLM:
$\text{g}(\mu_i) = \text{log}(\mu_i) = \beta_0 + \beta_1*X1_i$
$E[Y_i|X_i] = \mu_i = \text{exp}(\beta_0 + \...
0
votes
0
answers
43
views
Post-hoc test for a glmer model fitting a paired sample, repeated measures, poison distribution
I have paired sample dataset where the same sample received Treatment('Drug A','Placebo') on different days and i am comparing the effects of my response variable (i.e. Left_Pokes) per minute (i.e. ...
- 45
4
votes
1
answer
176
views
How to interpret the coefficients of Tweedie GLM with log link?
I'm trying to model cost data which have 0s. It seems that gamma is not an appropriate distribution and zero inflated gamma seems to be a bit of an overkill, but Tweedie seems to be appropriate with ...
- 71
2
votes
1
answer
144
views
Negative Binomial Regression Model - Effect of Removing Significant Covariates
I have been dabbling in NB regression for less than a year now. I have applied the well known g.o.f. tests. Lately I started using the Conditional Moment (CM) test, described in Cameron and Trivedi ...
- 21
0
votes
0
answers
27
views
regression model for zero-inflated pseudo-continuous outcomes with negative values and practical solution in R?
I am looking high and low for how to model data where the outcome is pseudo-continuous (change in questionnaire scores) that can be negative and has a lot of "true" zeroes. I see talk about ...
- 31
0
votes
0
answers
16
views
How to plot relationship between each predictor and response in GLM with multiple predictor
I have a structural equation like this:
So I contruct two model which follow the formula of:
"B ~ A" and "C ~ A + B"
The parameter for "B ~ A" will be βB0, βB1, σB
The ...
0
votes
0
answers
11
views
Changing loss function for mgcv gam
I am currently using the mgcv GAM-package and would like to experiment with other loss-functions. Specifically I am interested in trying out a loss function, which minizes the relative RMSE rather ...
1
vote
1
answer
63
views
Binomial regression in R: lm() with logit, vs glm() with family=binomial
I am currently learning about GLMs. Suppose my predictor variable generates a number of "successes" and "failures" as the response variable. Naturally I want to model the ...
0
votes
0
answers
53
views
Meta-analysis of three glm models with coefficient estimate
I have ran three simple logistic regression models:
glm(factor(response) ~ factor(sex) + age + gene expression,
data = data_2,
family = binomial)
Now I ...
- 1
1
vote
1
answer
27
views
Interpretation of estimates for a model with 2 categorial factors
I ran a model with 2 fixed factors that are categorial:
Ind: 10 levels
CRENEAU: 4 levels
For the parameter nu (I am running a <...
4
votes
1
answer
92
views
Is it appropriate to present predicted probabilities from emmeans for a mixed-effects binomial logistic regression?
I am trying to understand how to analyze data for a generalized mixed model (GLMM) with a binary response. I am interested in visualizing the predicted probabilities, as well as a measure of effect ...
- 43
0
votes
0
answers
25
views
Can expected variance of GLM be expressed through gradients under non-cannonical link functions?
In GLMs (generalized linear models), one can typically obtain an estimate of the expected variance of the response variable given the predictors as a transformation of the same parameter that defines ...
- 547
0
votes
2
answers
27
views
What to plot when main effect significant but interaction insignificant?
This may be a silly question but:
Given a regression model that includes both main effects and an interaction term
(Score of participants predicted by their attitude and motivation):
...
- 153
5
votes
2
answers
162
views
Nagelkerke's Pseudo-R² is negative? How to fix this?
I am working on my Master's Thesis and I was fitting various glms on my data; and since I can't calculate adjusted R² values for my models, I opted for Nagelkerke's pseudo R². I used the rcompanion ...
- 51
1
vote
1
answer
18
views
Why does my GLARMA model fail to converge when I exclude the independent variables, while a similarly structured one does not?
For purposes of validating a forecast model, I'd like to compare a GLARMA model that I developed to a null model that includes the same autocorrelation effects but lacks the environmental data. GLARMA ...
- 255
0
votes
0
answers
29
views
Opposite results from different modelling decisions in logistic regression
I'm trying to test the significance of 2 independent variables (b & c) in a regression model, whose values depend on another independent variable a. To make this easier to understand, let's say we'...
0
votes
1
answer
37
views
Using a General Linear Model in SPSS to analyze two variables over time? [closed]
I'm struggling a little with an analysis I need to run in SPSS. I am analyzing the effect of openness on self-esteem over three timepoints. I want to make sure time is included as a factor in order to ...
2
votes
1
answer
47
views
Is nesting needed if pot is the experimental unit?
I am analyzing an experiment where I grew two genotypes (native and invasive) of the same plant in one pot and evaluated growth over time. Half of the pots had two plants of the same genotype (native-...
1
vote
0
answers
29
views
Survey Package in R - svyglm with quasipoission link function for binary outcomes data, yielding relative risk? [closed]
I'm planning to run a regression analysis to estimate [adjusted] relative risk of a binary outcome data (essentially, analyzing the relative risk or risk ratio of an event happening), using a number ...
- 71
1
vote
0
answers
52
views
Python statsmodels GLM - log likelihood of null model
I have an issue when calculating log-likelihood for null model to double-check GLMResults.llnull parameter:
https://www.statsmodels.org/devel/generated/statsmodels.genmod.generalized_linear_model....
- 111
1
vote
1
answer
35
views
interpretation of GLM output
I want to get verified if I can say like this.
here is the GLM output:
...
- 53