semweb2slide48

From CommerceNet Wiki

Jump to: navigation, search

Image:semweb2.slide.48.png

Prev Slide Index Next Slide

The final concept for making the Web more semantic is vertical search.

Horizontal search engines such as Google work well for finding popular web sites. However, when the site you want is ranked 13,000 out of 147,000, or when the answer to your question must be compiled from information spread over many sites, current search engines are found wanting.

Vertical search engines overcome these limitations by using models of the user, the domain, and the available information sources to determine where and how to look. Vertical search often exploits deep web sources that aren’t accessible to Google’s spiders, either because they are proprietary or require specialized access protocols (e.g., SQL queries, filling out forms). The raw information retrieved from these sources is initially organized into data- or knowledge-bases. Domain and user models are then applied to connect the dots and return useful answers.

Let’s look at three examples.

Retrieved from "http://wiki.commerce.net/wiki/semweb2slide48"
Personal tools