Hyper-parameter is a type of parameter for a machine learning model whose value is set before the model training process starts.
Most of the algorithm including Logistic Regression deals with useful hyper parameters.
In this post we are going to discuss about the sklearn implementation of hyper-parameters for Logistic Regression. Below is the list of top hyper-parameters for Logistic regression.
- Penalty: This hyper-parameter is used to specify the type of normalization used. Few of the values for this hyper-parameter can be l1, l2 or none. The default value is l2.
- Inverse of regularization: This hyper-parameter is denoted as C. Smaller values of this hyper-parameter indicates a stronger regularization. Default value is 1.0
- Random state : random_state is the seed used by the random number generator. Default value is None.
- Solver: This indicates which algorithm to use in the optimization problem. Default value is lbfgs. other possible values are newton-cg, liblinear, sag, saga.
- Max iter : max_iter represents maximum number of iterations taken for the solvers to converge a training process.
I hope you enjoyed reading this post. Happy Learning !!