Code for "Unified Dual-Intent Translation for Joint Modeling of Search and Recommendation" in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2024)

Data

Source Data: download MT-Small data for UDITSR from https://drive.google.com/drive/folders/1Q7N5ppK7xvSud_rLGuYvoS-sgu8d0ZQS

Structure:

code

models

BaseP.py

UDITSR.py

conf.py

main.py

utils.py

UDITSR_data

train_data.csv

valid_data.csv

test_data.csv

data format:

• train_data

  user	query	Item	label
  0	[]	0	1
  1	[]	1	1
  2	[6]	2	1

• test/valid data. For each ground truth valid/test data, we randomly select 99 item that the user has not interacted with, serving as negative samples. Records sharing the same 'sample num' indicate that they are either a single positive sample or the negative samples derived based on the positive sample.

  sample_num	user	query	Item	label
  1	7	[42,6]	7	1
  2	14	[]	3461	0
  3	19	[]	678	0

cd code
python3 main.py