Answers, not links: extracting tips from yahoo! answers to address how-to web queries

  • Authors:
  • Ingmar Weber;Antti Ukkonen;Aris Gionis

  • Affiliations:
  • Yahoo! Research Barcelona, Barcelona, Spain;Yahoo! Research Barcelona, Barcelona, Spain;Yahoo! Research Barcelona, Barcelona, Spain

  • Venue:
  • Proceedings of the fifth ACM international conference on Web search and data mining
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

We investigate the problem of mining "tips" from Yahoo! Answers and displaying those tips in response to related web queries. Here, a "tip" is a short, concrete and self-contained bit of non-obvious advice such as "To zest a lime if you don't have a zester : use a cheese grater." First, we estimate the volume of web queries with "how-to" intent, which could be potentially addressed by a tip. Second, we analyze how to detect such queries automatically without solely relying on literal "how to *" patterns. Third, we describe how to derive potential tips automatically from Yahoo! Answers, and we develop machine-learning techniques to remove low-quality tips. Finally, we discuss how to match web queries with "how-to" intent to tips. We evaluate both the quality of these direct displays as well as the size of the query volume that can be addressed by serving tips.