标题: Probit vs. Logit? Need quick examples to remember! [打印本页] 作者: Matori 时间: 2013-4-6 03:31 标题: Probit vs. Logit? Need quick examples to remember!
I’ll provide some for time series vs. multiple regression. But I need
Time series: Autoregressive (rely on its own past), x-variable values are simply progressions in time (1..2..3..etc) stuck to a trend coefficient.
Examples: Sales levels are determine by their own lagged values. y = 4 + 0.09x (x=1,2,3..etc)
Multiple regression: Multiple, unique traits as x-variables.
Example: Sales levels are determined by GDP growth, population growth in the area over time.
Multiple regression uses unique variables over time, time series uses the y-variable as its independent variable, or just an arbitrary slope coefficient with no economic name of meaning.
Please correct me if wrong, and contribute!作者: former 时间: 2013-4-6 03:35
What does this have to do with Probit and Logit models? They are QUALITATIVE models, while the models/examples you described above are QUANTITATIVE.作者: thommo77 时间: 2013-4-6 03:39
Uh calm down…thought I’d just provide some other examples for people concerning time series and regression. But I would like an easy way to remember probit/logit, other than just sort of knowing they help predict bankruptcy/rank observations/are “discriminant” models.
Thanks!
