|
|
|
hakia is a semantic search engine that brings relevant results based on concept match rather than keyword match or popularity ranking. A detailed presentation can be found in hakia-Lab. There are two main components of hakia as explained below:
QDEX Infrastructure
hakia's designers broke from decades-old indexing method and built a more advanced system called QDEX (stands for Query Detection and Extraction) to enable semantic analysis of Web pages, and "meaning-based" search. See hakia-Lab for QDEX details.
SemanticRank Algorithm
SemanticRank algorithm is comprised of
innovative solutions from the disciplines of Ontological Semantics, Fuzzy Logic, Computational Linguistics, and Mathematics. See hakia-Lab for SemanticRank details.
Semantic technology enables many capabilities that are new to the search world. These are explained below.
CATEGORIZATION
For short queries such as cancer or Winston Churchill, hakia is able to present search results in a categorized form
to provide meaningful variations of the subject. Each gallery has 10 categories on the average, that is equal to running 10 queries in conventional search engines.
PARALLELISM
A very important capability for search, parallelism allows the hakia engine to bring results using equivalent terms such as "treat=cure" or "kill=murder". These associations are deployed with proper disambiguation methods.
GENERALIZATION
Most queries to search engines require generalization capability. For example, in the query "what car has the best mileage" the term "car" should be recognized as a general term for brands like Toyota, Ford, Chrysler, VolksWagen. Thus, the result should include the brand name instead of the word "car."
HIGHLIGHTING
For complex and longer queries, hakia highlights relevant phrases or sentences that best correspond to the meaning match of your query. Try
Why did Enron collapse? In most cases, you do not have to open the documents to see the quality of the results - a key to saving search time!
UNINTERRUPTED TEXT
hakia often displays uninterrupted sections of Web pages in search results that provide a full point of view of the content. Presenting complete snippets enables you to evaluate the search results instantly and saves you significant search time.
DIALOGUE
hakia points out good answers in a dialogue mode in addition to making suggestions, correcting spelling errors, and listing related hakia Galleries. This hakia capability continues to evolve. Try
What is the most common type of volcano?
COMPRESSED MEANING REPRESENTATION (CMR)
CMR is the capability of abstraction given any content. The level of abstraction is controllable. This utility is used in the back-end operations during QDEXing. CMR will also be available via technology licensing and syndication services.
|
|
|
|
|
