search

recommenders-team / recommenders

Best Practices on Recommendation Systems
https://recommenders-team.github.io/recommenders/intro.html
MIT License
18.89k stars 3.07k forks source link

[FEATURE] Data preprocessing and scoring for top-N recommendation on movielens for xdeepFM #585

Closed yexing99 closed 5 years ago

yexing99 commented 5 years ago

The followings need to be finished:

  1. For scoring the topN recommendation, the data needs to be pre-proccessed to have the ranking pool.
  2. Add scoring components for topN recommendation using reco repo
yexing99 commented 5 years ago

Finished

yueguoguo commented 5 years ago

Finished

@yexing99 does it mean it is resolved?

loomlike commented 5 years ago

Regarding generating a ranking pool (all user-item pairs or except the pairs in the training set if we want to remove 'seen' items), it is already there under reco_utils/dataset/pandas_df_utils.py. But if need to be a specific form other than our pd.DataFrames, feel free to add util functions.

yexing99 commented 5 years ago

The work is already done.. Your script does not apply here.

Get Outlook for Androidhttps://aka.ms/ghei36

From: Jun Ki Min notifications@github.com Sent: Friday, March 1, 2019 2:33:48 PM To: Microsoft/Recommenders Cc: yexing99; Mention Subject: Re: [Microsoft/Recommenders] [FEATURE] Data preprocessing and scoring for top-N recommendation on movielens for xdeepFM (#585)

Regarding generating a ranking pool (all user-item pairs or except the pairs in the training set if we want to remove 'seen' items), it is already there under reco_utils/dataset/pandas_df_utils.py

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/Microsoft/Recommenders/issues/585#issuecomment-468784727, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AYSLIKsFhh1pwmFeHoAwCMy7yh9NVOo_ks5vSYCcgaJpZM4bW3k3.

yexing99 commented 5 years ago

All the work has been finished. It is under my branch. I opened this ticket just for purpose of reporting progress as I should do it previously.

yueguoguo commented 5 years ago

@yexing99 I am closing this issue now unless there is anything needed to invoke team's attention. Do you want to merge the code that resolves this issue into the staging branch? Or you will wait till later when everything is finished?