How fit pairwise ranking models in xgBoost?

Question

As far as I know, to train learning to rank models, you need to have three things in the dataset:

• label or relevance
• group or query id
• feature vector

For example, the Microsoft Learning to Rank dataset uses this format (label, group id and features).

1 qid:10 1:0.031310 2:0.666667 ...
0 qid:10 1:0.078682 2:0.166667 ...

I am trying out xgBoost that utilizes GBMs to do pairwise ranking. They have an example for a ranking task that uses the C++ program to learn on the Microsoft dataset like above.

However, I am using their Python wrapper and cannot seem to find where I can input the group id (qid above). I can train the model using just the features and relevance scores but I feel like I am missing something.

Here is a sample script.

gbm = XGBRegressor(objective="rank:pairwise")

X =  np.random.normal(0, 1, 1000).reshape(100, 10)
y = np.random.randint(0, 5, 100)

gbm.fit(X, y) ### --- no group id needed???

print gbm.predict(X)

# should be in reverse order of relevance score
print y[gbm.predict_proba(X)[:, 1].argsort()][::-1]

Answer 1

According to the XGBoost documentation, XGboost expects:

• the examples of a same group to be consecutive examples,
• a list with the size of each group (which you can set with set_group method of DMatrix in Python).

Answer 2

set_group is very important to ranking, because only the scores in one group are comparable. You can sort data according to their scores in their own group.

For easy ranking, you can use my xgboostExtension.