Webicina search: Interview about Semantic indexing August 30, 2011

Now that Webicina.com has a new design and a brand new search engine, I thought it would be useful to show the basic features and I also publish an interview with Endre Jóföldi, CEO at WebLib LLC, the company behind the search engine of Webicina.com.

If you do a search for diabetes, you will see

• whether Webicina has a diabetes resource (a blog, podcast, Facebook group, Twitter user, etc), collection (Diabetes and Web 2.0) or sub-section (Diabetes Mobile Apps).
• You can narrow the search by “social media collection” (e.g. a diabetes resource in the asthma collection), “curated dynamic news category” (e.g. news categories featuring diabetes resources), “resource type” (RSS, resource, subsection or collection) or “languages”.

• Please tell us more about WebLib! What kind of projects are you famous for?

Generally we are building intelligent search solutions for our customers like the NIH or Vanderbilt University Biomedical Library. Behind the scenes we are also working on a semantic web knowledge base to improve our core search engine.

• What is the engine behind the Webicina search function like?

Our PolySearch enterprise search engine is a best-of-class domain independent semantic indexing and search solution built on top of the open source Solr/Lucene enterprise search platform. We have built several different search solutions based on this engine, which are all different in some sense based on the different needs of our customers.
PolySearch uses grammatical tools to find different forms of words and also uses our SearchComplete to offer health related search queries when the user typing his/her queries. It also utilize our medical spell checker to help correcting typos.

• How does clustering search help us find exactly what we need?

We face the problem of information overload after doing searches. For example if you search for cancer on Webicina  you will get more then 40 thousand results. Our relevance algorithm floats up the most important Webicina resources, however if you want to look through the tons of RSS results, it definitely helps to filter it down to your area of interest. Like you will see only 479 results in the Nutrition category. This is a much more friendly number.

• How difficult was it to develop a search engine for Webicina.com?

We tried to make a very fast engine, what helps Webicina visitors to find all the information Webicina treasures. It is also handling the different resource types and languages so we think it can really improve the user experience.  More than that we have still plans how to make it even better, so it is going to be a longer process, where we want to use the data how the visitors are going to use this search service.