Site Search that Leverage Content Structure for Relevancy

When a search engine searches the Web, it has to analyse information in an HTML format, which provides limited clues to the importance of different page components. However, the System 7 site search mechanism has direct access to the content of the online publications it powers, which is stored in a highly structured format. The System 7 site search mechanism uses its understanding of the structure of its content to increase the relevancy of the search results that it returns.

When an end-user of a System 7 publication performs a search, the site search mechanism examines indexes that are automatically generated on the different components of the content structure. From each index it obtains a relevancy indicator for an item, which it then combines with the other relevancy indicators using intelligent heuristics to create the best possible overall relevancy indicator.

For example, if an end-user searches for a term such as "income tax" in a publication, the System 7 site search mechanism can examine indexes on the article titles, the article leader sentences and the body text. When combining the different relevancy indicators for each item in order to create a single relevancy indicator that can be used to rank the item within the search results, the system is able to place greater emphasis on the indicator for the article leader sentence than the article body, and greater emphasis still on the indicator for the article title. In this way the relevance of different content items is accurately calculated, and effective search results returned to the user.

End-users increasingly rank site search amongst the most important features, and System 7 excels in this area.