Upload
anjan-goswami
View
95
Download
3
Embed Size (px)
Citation preview
Topic Models Based Understanding of Supply and Demand Side of an eCommerce Engine
Anjan Goswami and Wei Han@WalmartLabs Search Science
Generative Process
• Each topic is a distribution over words• Each document is a random mixture of corpus-
wide topics• Each word is drawn from one of those topics
• Our goal is to infer the underlying topic structure– What are the topics?– How are the documents divided according to those
topics?
Topic Model on Top Queries by Traffic
• Query is a short document, and not stable to build topic models
• Solution– Cluster queries using items
Topic Model on Aggregated Queries by Item
Xbox One Assassin's
Creed Unity Bundle
xbox one
xbox xbox one bundle
xbox one console
Results
X1 X2 X3 X4 X5 X6 X7 X8
1 printer ammo microwave bed bike pants vacuum curtains
2 hp diapers oven mattress bikes womens cleaner blinds
3 desk coffee gym twin xbox mens carpet curtain
4 ink pampers modem beds 360 women shark window
5 printers 22 treadmill foam games jeans bissell panel
6 computer ammunition paper queen console men hoover blackout
7 canon size router bunk bicycle hanes cleaners sheer
8 wireless maker toaster full pokemon women's steam drapes
9 dresserammunitinon wii frame cruiser danskin refrigerator mini
10 cartridge 3ds amiibo size ps4 sweatpants floor eclipse
20 topics (frequency and uniqueness)
tablet
dvd
table
tv
storagephone
baby
sofa
baby
laptop
bedding
kitchen
vacuum
clothing
other
bed
home
other
printer
curtain
Prediction On Search Queries
tablet
dvd
table
tv
storagephone
baby
sofa
baby
laptop
bedding
kitchen
vacuum
clothing
other
bed
home
other
printer
curtain
For each category, discover the gap between demand and supply (Simulated Data)
• Categories:– Category 1– Category 2– Category 3
References
• D. Blei. Probabilistic topic models. Communications of the ACM, 55(4):77–84, 2012.
• A. Chaney and D. Blei. Visualizing topic models. International AAAI Conference on Social Media and Weblogs, 2012.