Open fernandobperezm opened 7 months ago
Thanks for your kind attention!
Please refer to Figure 3 in the referenced paper. The prefixes a and b indicate two different datasets, RL4RS-Slate and RL4RS-SeqSlate.
Please refer to Table 2 in the referenced paper. The suffix _rl means the separated data before RL deployment. The suffix _sl means the separated data after RL deployment.
No, users' unique numeric identifiers are not provided. Please use session_id instead.
Please refer to the exposed_items column.
I don't think so, as it is not designed for the user-rating matrix.
Please see the reproductions section of the README file.
7&8. Detailed instructions on how to create Slate and SeqSlate datasets can be found in the project's tutorial, accessible here: https://github.com/fuxiAIlab/RL4RS/blob/main/tutorial.ipynb
Hi!
First and foremost, thanks for your contribution.
I'm using this dataset in my research; however, I'm having troubles to use the dataset after reading the SIGIR paper "RL4RS: A Real-World Dataset for Reinforcement Learning based Recommender System" . I'm hoping you could answer the following questions:
a_
andb_
prefixes in the data files? e.g.,rl4rs_dataset_a_rl
vsrl4rs_dataset_b_rl
._rl
and_sl
suffixes in the data files? e.g.,rl4rs_dataset_a_rl
vsrl4rs_dataset_a_sl
..unique()
operation on theuser_protrait
column. However, I got way more unique strings than what is reported in Table 2.item_feature
column, how can I identify the item numerical identifier? The paper says that the ID is inside this column but does not specify its position inside the array.Thanks in advance!