walterwsmf / tobit_wci

Tobit Model based on StatsModel with Confidence Interval in Python

tobit_wCI

Tobit Model based on StatsModel with Confidence Interval in Python

---

Developer:

Developer	e-mail
Walter Martins Filho	[email protected]

---

Credits:

I use the tobit likelihood and its seconde derivate from James Jensen, https://github.com/jamesdj/tobit, and just re-write the part for which package will implement the linear regression. In this case, I used the OLS routine from Statsmodel package.

Also, I add the routines to obtain the confidence interval from the Hessian matrix that came from scipy.optimize, which gave to us the maximum and minimum limits for our model.

---

Problems:

• tobit does not understand the values to censur when use apply_model from use.py

• it seems to have a problem how I include some dummie function on the problem.

---

Debug

Some warnings appears when we try to check errors on this code. But, it is only warnings:

1. warning in "tobit.py": No name 'log_ndtr' in module 'scipy.special'

   • Source: pylint-dev/pylint#2742