MirSiadaty.net

I wrote this relatively short document for users of search engines of MEDLINE (such as PubMed and BioMedLib), who don’t have or don’t want to hear about technical details of search and information retrieval, but want to make their search experience easier, faster, and better.

The eBook shows how you can have a much more focused and rewarding search without the gory detail of query syntax language, etc.

Here are headings of some of the sections:

• How to screen your search results faster
• How to minimize your “Absolute” and “Practical” misses
• Utilize the advanced search without being bugged by its gory details

Please feel free to post this eBook on your blog, email it, or link to it with whomever you believe will benefit from reading it. Thank you.

(http://bmlsearch.com/pdf/how_to_get_awful_pubmed_results.pdf)

Check out my weblog about biomedical information retrieval. I write it for the BioMedLib.com search engine.

http://www.bmlsearch.com/bmlblog/wordpress/

The blog is about applied aspects of “search” which directly relate to users.

The BioMedLib search engine is a free and open-access application which covers about 20 million publications (as of October 2010).

Since 2004 I have been working on improving retrieval of relevant publications from the MEDLINE® database.

MEDLINE is a list of about 20 million peer-reviewed biomedical research publications. New publications are added to it every day. There are many search engines that allow the user to search the information hidden within MEDLINE. For example, PubMed® and BioMedLib™ are such search engines. It is used by many health-care professionals, scientists, and some patients. It serves about 100 million queries per month.

If we could improve users’ search experience such that it took each user one minute less to locate the answer, then we would be saving more than 54 thousand hours of health-care professionals and scientists’ time every day. Such improvement should directly and indirectly help with more discoveries and better health care.

The goal of our project is to enable the user to create a better yet easier query and to do it faster; And then to be able to find the most relevant answers among the returned article, quickly and easily.

In technical terms, we are improving Sensitivity and Specificity of retrieval process simultaneously. Additionally we utilize text summarization; automated whole-corpus retagging of the 20 million articles, four times a year; automated re-creation of the biomedical knowledgebase which contains about 2.5 million concepts expressed by about 20 million synonyms, four times a year; and synchronized parallel real-time text processing on about 100 CPU cores.

Check out the current status of the product, the BioMedLib search engine, and tell me what you think.

Since this is my personal blog site, and assuming visitors would want to know about my background, here it is:

I started studying computer science (informatics) around 1988, the same year I entered the medical school of Tehran University. However the educational regulations at the time didn’t allow the students to graduate in these two fields simultaneously.  I took the MD.

In 2002, while working as the University of Minnesota employee, I completed Master of Science in BioStatistics.

So there you have it: M.D., M.S.

I didn’t get to complete my PhD from University of Virgina (dropped out in 2005). But the “search engine project” which started around the same time was at least as challenging and educational as a PhD in computer science.