Logistic Function Family

A Logistic Function Family is a function family that is restricted to logistic functions (of the form of the form $f(t,A,B,C)\equiv (C+A{e}^{-Bt}{)}^{-1}$, where $A,B,C$ are model parameters.)

• AKA: Maximum Entropy Model.
• Context:
  • It can be an input to a Logistic Regression Task (which results in a fitted logistic function).
  • It is in a Generative-Discriminative Relation with a Naive Bayes Model.
  • …
• Counter-Example(s):
  • a Probit Function Family.
  • a Naive Bayes Model.
  • a Linear Model.
• See: Logit Function.

---

---

References

2007

• http://www.nature.com/nrg/journal/v4/n9/glossary/nrg1155_glossary.html
  • LOGISTIC REGRESSION MODEL: A statistical model for the dependency of a binomial (two-class) phenotype on a number of risk factors. The probability, p, for one of the two phenotype states is expressed in the form of its logit, log(p/(1 – p)), which is assumed to be predicted by the linear combination (weighted sum) of the risk factors.

---