原文链接：https://xgboost.readthedocs.io/en/latest/prediction.html

There are a number of prediction functions in XGBoost with various parameters. This document attempts to clarify(澄清) some of confusions around prediction with a focus on the Python binding.

1、Prediction Options

There are a number of different prediction options for the xgboost.Booster.predict() method, ranging from pred_contribs to pred_leaf. The output shape depends on types of prediction. Also for multi-class classification problem, XGBoost builds one tree for each class and the trees for each class are called a “group” of trees, so output dimension may change due to used model. After 1.4 release, we added a new parameter called strict_shape, one can set it to True to indicate(表明) a more restricted output is desired. Assuming you are using xgboost.Booster, here is a list of possible returns:

• When using normal prediction with strict_shape set to True:
  Output is a 2-dim array with first dimension as rows and second as groups. For regression/survival/ranking/binary classification this is equivalent to a column vector with shape[1] == 1. But for multi-class with multi:softprob the number of columns equals to number of classes. If strict_shape is set to False then XGBoost might output 1 or 2 dim array.
  
• When using output_margin to avoid transformation and strict_shape is set to True:
  Similar to the previous case, output is a 2-dim array, except for that multi:softmax has equivalent output of multi:softprob due to dropped transformation. If strict shape is set to False then output can have 1 or 2 dim depending on used model.
  
• When using preds_contribs with strict_shape set to True:
  Output is a 3-dim array, with (rows, groups, columns + 1) as shape. Whether approx_contribs is used does not change the output shape. If the strict shape parameter is not set, it can be a 2 or 3 dimension array depending on whether multi-class model is being used.
  
• When using preds_interactions with strict_shape set to True:
  Output is a 4-dim array, with (rows, groups, columns + 1, columns + 1) as shape. Like the predict contribution case, whether approx_contribs is used does not change the output shape. If strict shape is set to False, it can have 3 or 4 dims depending on the underlying model.
  
• When using pred_leaf with strict_shape set to True:
  Output is a 4-dim array with (n_samples, n_iterations, n_classes, n_trees_in_forest) as shape. n_trees_in_forest is specified by the numb_parallel_tree during training. When strict shape is set to False, output is a 2-dim array with last 3 dims concatenated into 1. When using apply method in scikit learn interface, this is set to False by default.
  

Other than these prediction types, there’s also a parameter called iteration_range, which is similar to model slicing. But instead of actually splitting up the model into multiple stacks, it simply returns the prediction formed by the trees within range. Number of trees created in each iteration eqauls to treesi=num_class×num_parallel_treetreesi=num_class×num_parallel_tree. So if you are training a boosted random forest with size of 4, on the 3-class classification dataset, and want to use the first 2 iterations of trees for prediction, you need to provide iteration_range=(0, 2). Then the first 2×3×42×3×4 trees will be used in this prediction.