Validating a model with cross validation (logistic regression). What is Grid search for?

Question

scores = cross_val_score(logmodel,y_test.astype(float).reshape(-1, 1),predictions.reshape(-1, 1),
                        scoring="neg_mean_squared_error",cv=10)


log_rmse_scores = np.sqrt(-scores)

def display_scores(scores):
    print("Scores: ", scores)
    print("Mean: ",scores.mean())
    print("Standard Deviation: ", scores.std())


display_scores(log_rmse_scores)

Exit:

Scores:  [0.02972702 0.02972702 0.02972702 0.02972702 0.02972702 0.02972702
 0.02972775 0.02972775 0.02972775 0.02972775]
Mean:  0.029727313479823336
Standard Deviation:  3.574977450912211e-07

From what I understand, this is still not the best model: I need to "Tunar" the model seeking the best hyperparameters with Grid Search, that’s right?

What are these hyperparameters in logistic regression? At the end of the process, what will be the model?

I don’t understand what I get in the end.

Answer 1

In logistic regression there are no hyperparameters for join unless it is a logistical regression with regularisation.

Another thing that might be tunada in logistic regression are the variables (columns of table X) that enter the model. From what I understand from your code, table X has only 1 column, so you wouldn’t have what join in this case.