It is a commonly used statistical empirical analysis method in sociology, biostatistics, clinic, quantitative psychology, econometrics and marketing. Logit model is the earliest discrete selection model and the most widely used model at present. Logit model was first derived by Luce( 1959) according to IIA characteristics. Marschark( 1960) proves the consistency between Logit model and maximum utility theory.
The wide application of Logit model is mainly due to its advantages such as probability expression, fast solution speed and convenient application.
Extended data
According to the IIA characteristics of Logit model, the decrease or increase of selection branches does not affect the proportion of selection probability in other choices. Therefore, the selected branch that needs to be removed can be directly removed from the model, or the newly added selected branch can be added to the model for direct prediction.
Different from probability, a very important feature of Logit is that there is no upper and lower limit, which brings great convenience to modeling.
Logit model can overcome the defects of the model in predicting prior events to a certain extent, and integrate the advantages of FR probability analysis method in FR model and signal analysis method in KLR model. However, it is only based on several major financial assets or economic indicators, such as interest rate and exchange rate, which is different from the general early warning of currency crisis.
Baidu encyclopedia -logit model