How to make logistic regression model that obtained from literature review?

In case I know the full formula of the multivariate logistic regression model (such as exactly number of estimated intercept and variable coefficients of model parameters and the predictor variables) from the literature review, how can I make a model function follow formula to implement in the future?

for example, I know the value of b0, b1, ... and know what the predictors (x1, x2, ...) is

probability_y = exp(Z) / (1+exp(Z))

Z = b0 + b1x1 + b2x2 + b3x3 + b4x4 + ...

Answer 1

[Krishna]

Hello Kanokkwan,

It seems like you know the number of features as well as the value of weights you want to use in logistic regression. Now you want to train the network using that features and pre trained weights. For this you can use transfer learning. Please go through this documentation to learn more about transfer learning:

https://in.mathworks.com/discovery/transfer-learning.html

Also, logistic regression follows the same formula provided by you. So if you want to train the network from scratch use built in function provided by MATLAB. Please go through this documentation to learn more about logistic regression:

https://in.mathworks.com/help/stats/nonlinear-logistic-regression.html