Probit model

In statistics, the probit model is a specification of a generalized linear model and uses a probit link component. Probit models were introduced by Chester Bliss.

Application

Probit models are used as logit models to binary target variable ( 0/1) to model, ie, quantities that are on the lowest scale level. Examples: "Buy a product - Yes / No", " Can divorce - Yes / No", " Does High School - Yes / No".

Definition and estimation

Because the response is a series of binomial values ​​, the likelihood of adoption is subject, that it follows a binomial distribution. Be the Response and the vector of explanatory variables. The probit model has the assumption that applies

The distribution function for a standard normal distribution respectively. The parameters are typically estimated using the maximum likelihood method.

The probit model can be obtained by a simple Latent variable model. Assuming that

Being, and that is an indicator of whether the latent variable is positive:

Then it can be shown that the following equation is satisfied:

• Regression model
• Latent Variable Model
• Econometrics