164k views
2 votes
Select a distribution and link function for Continuous Positive Target (GLM)

1 Answer

4 votes

Final answer:

To model Continuous Positive Target in a GLM, the Gamma and Inverse Gaussian distributions can be used, along with the log link function.

Step-by-step explanation:

When it comes to selecting a distribution and link function for Continuous Positive Target in a Generalized Linear Model (GLM), one commonly used distribution is the Gamma distribution. The Gamma distribution is suitable for modeling continuous positive variables with a skewed distribution, such as response times or quantities.

As for the link function, a popular choice for continuous positive targets is the log link function. The log link function relates the expected value of the response variable to a linear combination of the predictors, with the log transformation ensuring that the predicted values are positive.

Another distribution that can be used for Continuous Positive Target in a GLM is the Inverse Gaussian distribution. The Inverse Gaussian distribution is often employed when dealing with data that has a long tail and may exhibit skewness.

User Scott Helme
by
8.1k points

Related questions

asked Dec 5, 2024 118k views
Dsollen asked Dec 5, 2024
by Dsollen
8.1k points
1 answer
4 votes
118k views
1 answer
5 votes
79.6k views