-
Notifications
You must be signed in to change notification settings - Fork 1
Logit regression
If one assumes that the probability is
Assume
Thus
Consequently,
Note that:
-
$v_{ij}$ is a linear combination of$X_{ijp}$ with weights$β_p$ as logit model parameters. - the odds ratio
$P(i \to j) \over P(i \to j')$ of choice$j$ against alternative$j′$ is equal to${w_{ij} \over w_{ij'}} = exp( v_{ij} - v_{ij'} ) = exp \sum\limits_{p} \beta_p \left( X_{ijp}- X_{ij'p} \right)$ - this formulation does not require a separate beta index (aka parameter space dimension) per alternative choice
$j$ for each exogenous variable.
Observed choices
Thus
Thus
with
The presented form
The latter specification can be reduced to the more generic form by:
- assigning a unique
$p$ to each$jq$ combination, represented by$A_{jq}^p$ . - defining
$X_{ij}^p := A_{jq}^p \times X_i^q$ for$j = 2..k$ , thus creating redundant and zero data values.
However, a generical model cannot be reduced to a specification with different
- limit the set of combinations of
$i$ and$j$ to the most probable or near$j$ 's for each$i$ and/or cluster the other$j$ 's. - use only a sample from the set of possible
$i$ 's. - support specific forms of data:
# | form | reduction | description |
---|---|---|---|
0 | general form of p factors specific for each i and j | ||
1 | q factors that vary with i but not with j. | ||
2 | p specific factors in simple multiplicative form | ||
3 | q factors that vary with j but not with i. | ||
4 | state constants Dj | ||
5 | state dependent intercept | ||
6 | usage of a recorded preference |
The
First order conditions, for each
Thus, for each
logit regression of rehousing logit_regression_of_rehousing "wikilink".
GeoDMS ©Object Vision BV. Source code distributed under GNU GPL-3. Documentation distributed under CC BY-SA 4.0.