CollaborativeFilteringModel train¶
-
train
(self, frame, source_column_name, dest_column_name, weight_column_name, max_steps=10, regularization=0.5, alpha=0.5, num_factors=3, use_implicit=False, num_user_blocks=2, num_item_block=3, target_rmse=0.05)¶ Collaborative filtering (ALS) model
Parameters: frame : Frame
source_column_name : unicode
source column name.
dest_column_name : unicode
destination column name.
weight_column_name : unicode
weight column name.
max_steps : int32 (default=10)
max number of super-steps (max iterations) before the algorithm terminates. Default = 10
regularization : float32 (default=0.5)
float value between 0 .. 1
alpha : float64 (default=0.5)
double value between 0 .. 1
num_factors : int32 (default=3)
number of the desired factors (rank)
use_implicit : bool (default=False)
use implicit preference
num_user_blocks : int32 (default=2)
number of user blocks
num_item_block : int32 (default=3)
number of item blocks
target_rmse : float64 (default=0.05)
target RMSE
Examples
See :doc: ‘here <new>’ for examples.