Stack Overflow
Stack Overflow

Questions tagged [tweedie]

A family of probability distributions, which is a special case of exponential dispersion models.

41 questions
8
votes
0 answers

Xgboost tweedie: Why is the formula to get the prediction from the link = exp(link)/ 2?

When I run a Tweedie GLM, one can get the prediction from the link by doing exp(link). To get the prediction for a Tweedie GLM, I get the prediction from the link by doing exp(link)/2. I don't understand why I need to divide by 2. Minimal…
Zoltan
  • 760
  • 4
  • 15
6
votes
1 answer

No access to "ldTweedie" function from mgcv when called from within another package

I am using mgcv package within my own package and so far everything worked nicely. However, I tried to introduce Tweedie family objects ("Tweedie()", "tw()") to my package, but I am getting an error when trying to use the latter. The function is as…
Daniel Piacek
  • 71
  • 4
5
votes
1 answer

LightGBM with Tweedie loss; I'm confused on the Gradient and Hessians used

I'm trying to figure out custom objective functions in LightGBM, and I figured a good place to start would be replicating the built-in functions. The equation LightGBM uses to calculate the Tweedie metric…
Sinnombre
  • 346
  • 1
  • 7
4
votes
1 answer

Transform SHAP values from raw to native units with lightgbm Tweedie objective?

The utility of Shapley Additive Explanations (SHAP values) is to understand how each feature contributes to a model's prediction. For some objectives, such as regression with RMSE as an objective function, SHAP values are in the native units of the…
kdoherty
  • 53
  • 7
4
votes
0 answers

Changing stepAIC to accommodate tweedie distributions

So I am trying to do a stepwise regression for a tweedie distribution. However, AIC is returned as NA by glm() if the family is tweedie, and this breaks the stepAIC command. I tried editting the code of the command to change extractAIC to…
J. Gursky
  • 121
  • 1
  • 9
4
votes
1 answer

Finite mixture of tweedie

I'm trying to estimate a finite mixture of tweedie (or compound Poisson-gamma) distributions. I have scoured any resources I can think of, without finding any resources on how to do this. I am currently trying to use the flexmix package in R writing…
masher
  • 41
  • 2
3
votes
1 answer

using xgboost to model a Tweedie regression

I'm trying to use xgboost to make a tweedie model, however I get an obscure error message. Here is a reproducible example: Preparing the data: library(xgboost) library(dplyr) set.seed(123) xx <- rpois(5000, 0.02) xx[xx>0] <- rgamma(sum(xx>0),…
Bastien
  • 3,007
  • 20
  • 38
3
votes
1 answer

GLM with Apache Spark 2.2.0 - Tweedie family default Link value

I am using spark 2.2.0 with python. I tried to figure out what is the default param of Link function Spark accepts in the GeneralizedLineraModel in case of Tweedie family. When I look to documentation…
Antoine Ly
  • 33
  • 1
  • 7
2
votes
0 answers

TweedieLoss in Pytorch Forecasting Model

I am currently implementing a demand forecasting solution using Deep Learning in Pytorch with multihorizon output, and I am experimenting with Pytorch forecasting (PTF) models (N-Beats, N-hits, DeepAR, etc), using TimeSeriesDataset and DataLoader…
George Gousios
  • 56
  • 5
2
votes
1 answer

Mean Residual Deviance Formula in H2O

I'm trying to find out the exact formula used in H2O for the Mean Residual Deviance loss function for a Tweedie distribution. Or even, in general, what would be the mean residual deviance for a Tweedie distributed dependent variable? So far, I've…
vosirus
  • 113
  • 1
  • 6
2
votes
1 answer

Parameters of ligthgbm tweedie

I'm using the tweedie option for the objective parameter of the lightgbm package. According to wikipedia, the tweedie distribution has a free parameter p, but I can't seem to find the way to give it a value. Does anyone know how to do it? What is…
Jaime o.
  • 21
  • 3
2
votes
1 answer

Lightgbm with Tweedie

I'm trying to run lightgbm with a Tweedie distribution. I believe this code should be sufficient to see the…
Simon Todd
  • 156
  • 1
  • 10
2
votes
1 answer

model selection with tweedie distributions in MuMIn package R

I'm trying to do AICc model selection and model averaging with tweedie (compound Poisson) distributed data in R. I was working with the AICcmodavg R package with no success, then decided to try out the MuMIn package when I came across the…
RachelF
  • 21
  • 2
2
votes
1 answer

mgcv: How to do stepwise regression with a Tweedie response model?

Does anyone have an idea how to do stepwise regression with Tweedie in R? I found the mgcv package, which apparently treats the power parameter of Tweedie as yet another parameter to be estimated. This seems to improve on having to use…
Alp Can
  • 33
  • 1
  • 6
2
votes
1 answer

Package dglm in R

I am trying to fit a double glm in R using the dglm package. This is used in combination with the statmod package to use the tweedie model. A reproduction of the problem is: library(dglm) library(statmod) p <- 1.5 y <- runif(10) x <-…
Acarbalacar
  • 714
  • 2
  • 7
  • 19
1
2 3