Page 1 of 1

Serendipity in research

PostPosted: 30 Nov 2016 12:45
by Richard
I'm working on some ideas about the importance of serendipity in research and came across serendipomatic.org - a site where you paste in some text and it searches libraries for tangentially associated sources. This is supposed to help you find things you wouldn't otherwise find.

I put the blurb from my recently published book into this but what came out was pretty much complete rubbish (ships with the name 'X topic'). If you try it, can you find anything useful?

Re: Serendipity in research

PostPosted: 01 Dec 2016 10:15
by Nathan
I tried a couple different searches and had similar results. Apparently, instructed second language acquisition has something to do with grilling meat in Australia? This must be some really deep stuff on a level that I'm not at yet...
Or... complete rubbish.

I like the idea of serendipity in research though. This happened a lot to me in completing my MA research project. One thing led to another and now I find myself mining the topic with three or four possible papers to work with. For me, narrowing down is always much more of an issue than branching out and serendipity has a lot to do with it.

Re: Serendipity in research

PostPosted: 02 Dec 2016 15:36
by punjaporn
It achieves its goal of letting us know connections we never knew existed!!
What really surprised me is that the BNC has something to do with Sphinx. :mrgreen:

Re: Serendipity in research

PostPosted: 03 Dec 2016 13:47
by sgtowns
I put in the abstract of my most recent conference paper. The paper was on four different ways to use lexical cohesion: text evaluation, text segmentation, text summarization, and text criticism. But the search words chosen by serendipomatic.org were:

criticism, methods, text, purposes, surface, two, four, texts, analysis, actually, apparently, different, varied, studying, review

So right from the beginning, it is easy to see that the search will not return useful results for me because these are very general words. Can you imagine searching Google Scholar with "methods, purposes, analysis"? I'd guess that a fairly large percentage of papers ever published have these three words in them. And "two" and "four", or "text" and "texts" are not going to help much either.

On the other hand, I had the wonderful opportunity to read "Morphometric analysis of the fascicular organisation of the optic nerve" from the National Library of Serbia on Europeana written by Radunović Miroslav, Vitošević Zdravko, Ćetković Mila, Vuksanović-Božarić Aleksandra, Radojević Nemanja, and Radunović Miodrag. The abstract did in fact contain the matched words of: methods, surface, four, analysis, different, and varied.

Criticisms aside, this is a cool idea. Seems like they need to improve their word selection algorithm and perhaps get some better sources to search. But the idea is pretty neat.

Re: Serendipity in research

PostPosted: 03 Dec 2016 14:06
by sgtowns
I was just thinking a bit more about this idea of serendipity and "tangentially associated sources" that Richard was talking about in his first post. I subscribe to several Google Scholar email alerts on topics that are related to my research. So whenever new papers written on that topic are added to the Google Scholar index, then I get an email about them. This is sometimes better than just searching for the topic in Google Scholar because it's new/current research, and therefore it brings up papers that aren't going be on the first few pages of the Google Scholar results. The papers that it links to in an email aren't always completely relevant, but definitely some of them have been in very interesting tangentially associated fields.

To set up the email alerts, just search for a topic on Google Scholar, and on the bottom of the left side menu of the search results is a link for "Create Alert". Clicking there will show you some of the papers that would be sent to you. This list is much different than the original search -- the initial Google Scholar search for "writing quality" gave me the important, highly cited publications from 1981, 1986, 1999, etc. but the email alert will send me papers that are all from 2016. This Google Scholar feature has been really useful for me, so I highly recommend everyone set up a few alerts in whatever your topic is to help you "serendipitously" find new papers.

Re: Serendipity in research

PostPosted: 07 Dec 2016 09:49
by punjaporn
Thanks Stuart. As you suggested, I set up alerts for some topics on Google Scholar.
While playing around with this, I also find another feature useful - the alerts for Scholar’s new articles and new citations.

Mike Scott is a key person in Keyword analysis, so I searched for his profile on Google Scholar homepage, followed him, then set up alerts for his new citations. I’ve received “New citations to articles in Mike Scott’s profile” email with papers from many fields of studies, such as IT, text mining, communication science, medical sciences. I think those papers are interesting although some of them are not fully relevant.