hakia.com - The Semantic Search Engine
Hakia is an award winning, semantic search engine and one of the first major players in the it’s field. While the site is still in beta, my tests have shown some positive results… not perfect, but it’s definitely a move in the right direction.
Interestingly enough, when you do a search for “What is a semantic Search engine” in Google… it brings back Hakia.com as the first result. So, even based on that criteria, they must be doing something right.
Aren’t All Search Engines Semantic?
The simple answer to that question is; No.
Most search engines are still keyword based. This means that even if you ask Google or Yahoo! an actual question like “What is a Semantic Search Engine?”, the site will remove parts of the search it believes are unneeded and just perform a search on the keyword. In this case, it would remove “what is a” and just do a search for “Semantic Search Engine”. Since it removed the parts that make it a question, Google thought I was looking for an actual search engine.
Hakia on the other hand will actually try to read and understand what you are saying by dissecting the sentence structure. So, by asking Hakia the same question “What is a Semantic Search Engine”, the program was able to determine that I am indeed looking for an answer to my question… not just looking for a keyword or web site.
This Isn’t a Popularity Contest.
Another feature that separates hakia from other search engines is the way it categorizes it’s results. Most search engines will order their results on what essentially becomes a popularity contest. Meaning, it not only searches for relevant sites to match your query, it also looks at incoming links, references on other sites, and other factors to determine a site’s rank. Granted, this is a good way to determine the reputation of a site, because after all… if sites are linking to them, it probably has some good information. On the downside, it creates a barrier to new sites even if they have good information.
What Hikia does different is that it determines the order based on three things: 1. The information comes from a credible source 2. Is recent information 3. The information is relevant to the query. The real strength of this engine however comes from #3… relevancy to the query. Exactly how it does this is a very complicated procedure and one would take too long to go into detail here. Here’s an excerpt taken from their labs page that details how it works:
…the QDEX system analyzes the entire content of a Web page (including HTML). Then, the QDEX algorithm extracts all possible queries that can be asked to this content, at various lengths and forms. These queries (sequences) become gateways to the originating documents, paragraphs and sentences during the retrieval mode. Note that this is done off-line before any actual query is received from a user.
Search Results in Three Parts.
Ok, so now that we know a little bit about the technology and processes behind the search engine, let’s talk briefly about the user interface. Hakia divides it’s main search results page into three separate sections:
- 1. Relevant, Credible, and Fresh Results
- 2. Images
- 3. The Rest of the Results
The main thing you’ll notice when you reach the search results page is that the division between the quality results on the top and regular search results on the bottom, separated by a line of images.
Now, I do have to say that this approach can be extremely useful if you’re looking for information on health topics where it’s important to have credible references. But, if you’re looking for something simple like “Strollers”, the first “credible” results are news stories and government reports… which can be a little annoying if you’re just looking to browse for strollers and have to scroll down before you see the “normal” search results.
Hakia also provides a few other search result layouts for discovery type searches and deep vertical searches. Each of these results really require their own separate post to do them justice, so we’ll leave that to a later post. But, I also strongly encourage you to explore all aspects of this site on your own.
Conclusion
Even though Hakia has taken a back seat to other semantic search engines in the news, it’s had a head start in the Semantic Search field and had a chance to really hone it’s indexing capabilities. If you’re looking to give a semantic search engine a try, I recommend adding this to your browser. (just don’t ask me how to correctly pronounce the name :)
August 18th, 2008 at 1:44 pm
[...] hakia.com - The Semantic Search Engine [...]