Tuesday, June 18, 2024

Microsoft: Noise Improves Search Results

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Incorporating data obtained from observing user behavior can significantly improve the ordering of search engine top results.

So-called “noisy behavior” — information on where users click, how long they stay on sites and whether and how they reformulate their queries — can improve search results by 31 percent, according to a new study from Microsoft Research.

The implications are potentially huge in combating click spam as well as improving the quality of search results.

In contrast to click fraud, where advertisers or syndicators click on each others’ ads in order to inflict higher costs on their competitors, click spam sites contain no useful information other than advertising.

Search engines are being attacked by businesses and consumer activists for failing to prevent both types of fraud.

Eliminating click-spam sites from top search results would be one way of addressing this increasingly vexing problem.

“By examining click-through and browsing patterns across a large number of users, we are able to learn a great deal about how people interact with search technologies and thereby improve our accuracy dramatically,” said Eugene Agichtein, study co-author and a researcher in the text mining, search and navigation group within Microsoft Research.

Agichtein and his team are presenting the papers at the annual ACM SIGR (Association for Computing Machinery’s Special Interest Group on Information Retrieval) conference in Seattle this week.

In one study of user behavior, the authors performed a large-scale evaluation over 3,000 queries and more than 12 million user interactions with a major search engine.

The authors wrote that their work “has many potential applications including click spam detection, search abuse detection, personalization and domain-specific ranking.”

The researchers took rankings obtained from a search engine query and re-ranked them based on algorithmically-derived interpretations of user behavior.

The authors also published a second study showing that search engines can improve rankings through accurate modeling and interpretation of user interactions.

Ben Edelman, an activist and researcher studying advertising fraud at Harvard University’s Department of Economics, saluted Microsoft for doing this type of research.

According to Edelman, click spam has been a thornier problem for Microsoft’s MSN search engine than for Google or Yahoo.

“But to their credit, they’re getting better every day,” said Edelman.

This article was first published on internetnews.com, a JupiterWeb site. To read the entire article, click here.

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles