I came across the following error while fitting a LightGBM classification model. All of the features in use are numerical with around 20% missing values. The target variable is binary. Could I know how this error (Check failed: best_split_info.left_count > 0) is resulted?
from lightgbm import LGBMClassifier
lgb_model = LGBMClassifier()
lgb_model.fit(train[predictors], train['is_booked'])