I have developed machine learning model (XGBClassifier) using different sets of variables, and both sets achieve an accuracy score of around 80%. However, one model uses 17 variables, while the other model uses 18 variables. I’m unsure which model to choose for deployment. Here are some factors I’m considering: