Ratios, differences, and link functions?

I recently learned that with Poisson regression, you can model rate difference by using an identity link function, or rate ratio by using a log link function. Does that work the same way with other generalized linear models? For example: If I fit the linear regression equation “height_in_cm = B0 + B1*gender” where gender is … Read more

LIME – local linear explanations for global black-box models work for LogisticRegression but not for RandomForest?

I’m trying to use LIME to locally explain black-box models predictions. It works by generating 5000 neighbors for an instance by perturbing the features. Then it does forward-selection by iteratively fitting a simple Ridge model over that local neighborhood, where the label is the black-box model prediction. After n iterations, the selected features are considered … Read more

Chi square and p value in GAM results in R

I would like to know how to interpret chi square and p values in GAM. As I understand p values basically tell us a story about the graphs – how low is the probability that there’s just a straight line in the graph, eg. s()=0. Am I correct? But what about the chi square? For … Read more

Interpretation of Fixed Interaction Effects in a Linear Mixed Model

I have a linear mixed model for predicting speed of a street crossing by a pedestrian with the factors VS (vehicle speed), GAP (selected time gap size for crossing), the interaction factors VSxGAP and a random intercept. How can I interpret parameter estimates for the interaction effects? I only have a random effect of the … Read more

Interpretation: Mediation with a squared regression model (U-shape)

I did a mediation analysis with Person-Organization Fit as IV, Organizational Attraction as Mediator and Organizational Choice Intention as DV. As the regression assumption ‘Linearity in the parameters’ was violated, I had to include Fit squared in addition to the IV Fit in the analysis. Unfortunately, I got results for the effects, that differ in … Read more

How to explain the gam plot to a non-statisticians

I have a GAM model to study the relationship between antibody titers at Cord (on the log scale) with the maternal antibody titers (on the log scale) and duration between vaccination and delivery. After doing some data exploratory steps, I decided to fit a gam model. This is my model: gam.PT.Cord <- gam(logcordpt ~ logmumpt … Read more

variable-specific output for MANOVA with error term?

I am analysing a multivariate data set of several measurements taken from a number of cells for individuals from two treatments (see example below) ID TREAT CELL HEAD_A_mm2 HEAD_L_mm HEAD_W_mm TAIL_mm TOTAL_L_mm 1 1541 C c1 5.967e-06 0.0032 0.0023 0.0336 0.0368 2 1541 C c2 6.038e-06 0.0037 0.0021 0.0325 0.0362 3 1541 C c3 5.537e-06 … Read more

Interpretation of incidence-rate ratios

So, I want to fit a random effects negative-binomial model. For such a model STATA can produce exponentiated coefficients. According to the help file such coefficients can be interpreted as incidence-rate ratios. Unfortunately I am not a native English speaker and I do not really understand what incidence-rate ratios are or how I could translate … Read more

2D rotation of standardized PCA, how to compute new eigen values, eigen vectors, and loadings ?

I want to apply a 2D rotation of a θ angle to my two first principal components of a PCA. What I understood from this post is that I have to apply a rotation matrix R : Rθ=(cosθ−sinθsinθcosθ) If the coordinates of my observations in the factorial plan are the matrix Y of dimension n∗2, … Read more