Dealing with e-commerce data in deep learning

Deep Learning
1

hey all,

I feel like I am kind of lost in dealing with e-commerce data. I wanted to deep-dive and understand the concept since there are so many things to do in this field. I thought I consumed some information at some point. Still, when I started the Kaggle - Expedia competition, I was kind of frozen since there are a significant number of id-based columns which contain many unique values in them. I couldn’t find any solution for how I should be leveraging them. I also tried to follow collaborative filtering in fast.ai docs but it looks like the data types differentiate from each other.

I would appreciate anything that would broaden my mind. Thanks in advance.

2

Hi,

Look up the course lesson on Tabular Classification which has tools to convert categorical data.

Also here…

1 Like
3

Thanks for sharing resources @meanpenguin but, I wasn’t truly looking for such sources. I am more into a bit more advanced stuff. For instance, adding additional features to collab_learner rather than just simply using only user_id, movie_id, and rating as in here.

Example dataset:

index user item rating age gender occupation movie_title 2 3 4 5 6 7 8 9 10 11 12 13 14
12323 428 338 4 28 M student Bean (1997) 0 0 0 0 0 1 0 0 0 0 0 0 0
3341 393 66 3 19 M student While You Were Sleeping (1995) 0 0 0 0 0 1 0 0 0 0 0 0 0
87579 445 844 2 21 M writer Freeway (1996) 0 0 0 0 0 0 1 0 0 0 0 0 0
46749 488 162 3 48 M technician On Golden Pond (1981) 0 0 0 0 0 0 0 0 1 0 0 0 0
35134 92 68 3 32 M entertainment Crow, The (1994) 0 1 0 0 0 0 0 0 0 0 0 0 0