Publication Type

Conference Proceeding Article

Version

publishedVersion

Publication Date

8-2019

Abstract

E-commerce taxonomy plays an essential role in online retail business. Existing taxonomy of e-commerce platformsorganizes items into an ontology structure. However, theontology-driven approach is subject to costly manual maintenance and often does not capture user’s search intention,particularly when user searches by her personalized needsrather than a universal definition of the items. Observingthat search queries can effectively express user’s intention,we present a novel large-Scale Hierarchical taxOnomy viagrAph based query coaLition (SHOAL) to bridge the gapbetween item taxonomy and user search intention. SHOALorganizes hundreds of millions of items into a hierarchicaltopic structure. Each topic that consists of a cluster of itemsdenotes a conceptual shopping scenario, and is tagged witheasy-to-interpret descriptions extracted from search queries.Furthermore, SHOAL establishes correlation between categories of ontology-driven taxonomy, and offers opportunitiesfor explainable recommendation. The feedback from domainexperts shows that SHOAL achieves a precision of 98% interms of placing items into the right topics, and the resultof an online A/B test demonstrates that SHOAL boosts theClick Through Rate (CTR) by 5%. SHOAL has been deployed in Alibaba and supports millions of searches for online shopping per day.

Discipline

Computer Engineering

Research Areas

Data Science and Engineering

Publication

Proceedings of the 45th International Conference on Very Large Data Bases, Los Angeles, California, 2019 August 26-30

Volume

12

Issue

12

First Page

1858

Last Page

1861

Identifier

10.14778/3352063.3352084

City or Country

Los Angeles

Additional URL

https://doi.org/10.14778/3352063.3352084

Download

Share

COinS