Posts Tagged ‘algorithm’
Search Engine Evolution
July 8th, 2009 | Author: Yohay Barsky CTO
Do we really need millions of “possibly related results” or just a few great results to answer our information needs?
Search engine quality is measured by two things: (1) “coverage power,” the number of scanned sites and the scanning frequency, and (2) ability to match the relevant results to each and every searcher’s information needs. Today, coverage power is extremely impressive; however, the problem of “search results mismatch” remains unsolved because of the Internet’s ever-growing complexity and the tremendous increase in the amount of information available.
The source of the “mismatch” may begin with the user’s inability to generate an accurate search query. According to research, the average user’s query length is 2-3 words . Google uses its famous PageRank algorithm to statistically enhance the relevancy of the results (delivering the most likely matches first). But the statistical heuristic algorithm is bounded in part by the user’s search input. Read the rest of this entry »
|