/DA-for-SBR

Data augmentation method for Session-based Recommendation

Primary LanguageJupyter Notebook

# DA-for-SBR

Data augmentation method for session-based recommendation

item :
FA = Frequency of appearance
HS = Highest similarity value

method1

method2

method3


Start

by running preprocessing.py, following folder hierarchy is made.

da_for_sbr(main folder)
  |-main.ipynb
  |-preprocessing.py
  |-simmetric.py
  |-utils.py
  |-narm_torch
  |-srgnn_torch
  |-exps
  |  |-experiment1
  |  |  |-yoochoose (train files, test files)
  |  |  |-diginetica (train files, test files)
  |  |  |-result_narm_yoochoose
  |  |  |  |-y064
  |  |  |  |-y128
  |  |  |  |-y256
  |  |  |  |-y512
  |  |  |-result_narm_diginetica
  |  |  |  |-d001
  |  |  |  |-d003
  |  |  |  |-d006
  |  |  |  |-d012
  |  |  |-result_srgnn_yoochoose
  |  |  |  |-y064
  |  |  |  |-y128
  |  |  |  |-y256
  |  |  |  |-y512
  |  |  |-result_srgnn_diginetica
  |  |  |  |-d001
  |  |  |  |-d003
  |  |  |  |-d006
  |  |  |  |-d012
  |  |-experiment(n)
  |  |-train_item_views.csv
  |  |-yoochoose-clicks-withHeader.dat\

data download

-train_item_views.csv (https://drive.google.com/file/d/14_Gej2IzMyIR6bVQ0O7mR34Ln20NYt4v/view?usp=sharing)

-yoochoose-clicks-withHeader.dat (https://drive.google.com/file/d/14YJ6Pntx3a2b9If9DgzF9Ax2Po9SUk_W/view?usp=sharing)