Could someone provide a general estimate of the average (time) complexity of the LambdaMART learning-to-rank algorithm?
A particular implementation of LambdaMART is known as XGBRanker. It uses XGBoosts as the underlying model to learn to predict relevance scores. Is it possible to derive a better estimate for this particular implementation?
Please, provide the necessary citations/sources on which the complexity estimation is based
EDIT: Resources on LambdaRANK and LambdaMART