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.)

Filter by
Sorted by
Tagged with
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 ...
user3256536's user avatar
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 [...
Jeffrey Royer's user avatar
0 votes
0 answers
51 views

Why does weights argument in glm change the outcome of lsd?

I have dataframe df. ...
scott.pilgrim.vs.r's user avatar
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 ...
user401679's user avatar
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'....
Noob29's user avatar
  • 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 ...
Franco Ferrante's user avatar
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 ...
Tom Wenseleers's user avatar
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 ...
Ian Holdroyd's user avatar
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_{...
Firebug's user avatar
  • 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 ...
Epimetheus's user avatar
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 ...
Carlos Ramos's user avatar
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 ...
Marco De Virgilis's user avatar
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: ...
Beo's user avatar
  • 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 ...
Kozolovska's user avatar
  • 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 ...
jay's user avatar
  • 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,...
No-one's user avatar
  • 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 ...
Ilan Halperin's user avatar
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)...
user400814's user avatar
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. ...
scott.pilgrim.vs.r's user avatar
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 ...
Meep's user avatar
  • 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 ...
LukasDphd's user avatar
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 ...
KeyPi's user avatar
  • 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&...
Javier Fajardo's user avatar
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 ...
C. Murtaugh's user avatar
-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 $0$. ...
melatonin15's user avatar
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 ...
Kiara H's user avatar
  • 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 ...
user395714's user avatar
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 ...
ppanko's user avatar
  • 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? ...
Nicolas Bourbaki's user avatar
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 + \...
stats_noob's user avatar
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. ...
JLit98's user avatar
  • 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 ...
user395714's user avatar
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 ...
Barton's user avatar
  • 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 ...
gj.'s user avatar
  • 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 ...
elainesun442's user avatar
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 ...
John Hundesen's user avatar
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 ...
cambridgecircus's user avatar
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 ...
Sally's user avatar
  • 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 <...
Marie Guittonneau's user avatar
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 ...
user398696's user avatar
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 ...
anymous.asker's user avatar
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): ...
a.henrietty's user avatar
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 ...
Pauline's user avatar
  • 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 ...
ohnoplus's user avatar
  • 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'...
Ruiming Ma's user avatar
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 ...
acesandstars's user avatar
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-...
Plants n bugs's user avatar
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 ...
user395714's user avatar
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....
Paweł Orliński's user avatar
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: ...
user398060's user avatar

1
2 3 4 5
91