I have been trying to get a working LightGBM model which I can train on my data, select the best performing model with highest f1 score and then use it obtain the f1 score on the testing data. However, all the material I have found online gives me errors.
Currently, what I am doing is after splitting the data and test set (70:30), and preparing cv folds, I prepare the recipe, then the tuning parameters etc., eventually examining the models for parameter selection using F1 scores.
I have data which I want to use to predict the loan_default status and whether someone will default or not, using 21 predictors in my data set, using downsampling.
Does anyone have a working script or helpful resources?
Thank you in advance.