Maximizing log-likelihood function of Binary Logistic Regression and estimating probabilities of the dependent variable equaling 1:

 

maximize

 logexp_sum(matrix_1)

Value:

 logistic(matrix_1)

 

OR (equivalently):

 

maximize

 logexp_sum(matrix_1)

 calculate

Point: point_problem_1

Value:

 logistic(matrix_1)

 

where

 

matrix_1:

 

x1

x2

x3

x4

scenario_benchmark

1

1

1

1

1

1

-4

5

8

-3

-2

2

-8

-4

1

4

-4

8

-3

6

-4

-9

-1

7

0

1

1

0

1

0

 

The matrix_1 is in file matrix_1.txt in PSG Matrix format (see, PSG Matrix in General (Text) Format).

 

Optimal point (estimates of regression coefficients):

 

Component_name

Value

x1

x2

x3

x4

6.69e+001

6.38e+001

-1.69e+001

-1.13e+001

 

Value of the logistic function (estimates of probabilities):

 

value

6.52e-009

1e+000

1e+000

5.27e-040

9.99e-001

1.65e-009

 

Files problem_logistic_regression_Calculate_test.txt, problem_logistic_regression_test.txt, and  matrix_1.txt are in folder .\Aorda\PSG\Examples\Functions\logistic_regression_test.