But today we are not satisfied with the quality of information unearthed by general purpose engines like Google, Altavista, Ask jeeves to name a few. The trend is to refine search based on a particular domain.
If we look at www.searchenginewatch.com — a one-stop window to search engine news, a variety of domain-specific engines which can of course be further bifurcated are available.
Domain-specific
We know that search engines comprises of crawlers or spiders which sift through the World Wide Web to fetch relevant information. But when we want to restrict our search to a particular domain, general purpose web-based engines aren’t efficient.
This resulted in the popularity of domain-specific engines to take roots.
But the limitation of domain specific engines is they consume more time and require human intervention to moderate the key words. They are also not completely automated.
A majority of domai-specific search engines use machine learning algorithms to filter out key words that are irrelevant for a particular domain under consideration.
Japanese study
But in an interesting study done by Satoshi Oyama, Takasha Kokubo and Toru Ishida of Kyoto University, Japan, they unearthed a unique code for specialised searching.
One of the researchers was searching for beef recipes on a popular Japanese search engine. He used the key word gyuniku (beef in Japanese) and only 15 out of 25 (60%) results were related to beef. He modified the query — gyuniku (beef) + shio (salt) and found 24 out of 25 (96%) results relevant.
Algorithm
Based on this idea, they designed an algorithm, which works in three broad stages:
1. User’s query will be accepted by the domain-specific engine (beef).
2. The search will be expanded by automatically inserting relevant domain specific key word (salt).
3. The engine sends the query to general purpose search engines for an exhaustive search to improve the hit ratio.
Some of the domain-specific search engines are available at www.searchenginewatch.com are — answers searching, computer search engines, domain searching, financial search engines, government search, legal search engines, medical search engines, science search engines, shopping search, travel search engines, broad category of search engines: