LIS Links

First and Largest Academic Social Network of LIS Professionals in India

May Artificial Intelligence help in library classification?

So many conferences and debates are taking place on role of Artificial Intelligence in LIS.  Can Artificial Intelligence help librarians in classifying the reading materials?

Views: 776

▶ Reply to This

Replies to This Forum

AJIT PRABHAKARANPermalink Reply by AJIT PRABHAKARAN on April 25, 2012 at 21:14

YES SIR.. U MIGHT HAVE HEARD ABOUT I-GOOGLE (INTELLIGENCE GOOGLE), WHICH IS SPREADINGS ITS ROOTS IN EDUCATIONAL SECTORS TOO..

SEMANTIC WEB IS SOMETHING WHICH ITSELF DEFINES AS "MEANING ABOUT DATA" THAT IS FROM SEARCHING DOCUMENTS FROM CONTEXT TO CONTENT...

Permalink Reply by Ritesh Kumar Tiwari on April 25, 2012 at 21:56

Dear Ajit Sir,

Many many thanks for your visionary reply. Sir, soul of my question is would artificial intelligence be able in classifying books or reading materials according to any classification scheme and decrease the work load of classifier?

Permalink Reply by Digambar Bhaurao Waghmare on April 25, 2012 at 22:26

sir

your discussion topic is very interested I think someone should work on this topic 

Thanking you,

Digambar B Waghmare (Librarian)

SInhgad College of Engineering, Pandharpur (Maharashtra)

Permalink Reply by R Sunil Kumar on April 25, 2012 at 22:59

Yes it will....as Mr. Ajit had said that google is already doing first level of classification of data and creating metadata... time is very close when some programmer will create a program.

Program will first index the book and then sum up number of times a particular word is repeated and then find the relationship between different words and generate a classification number based on DDC/UDC/LCSH.

And it will do on all books scanned book with OCR, websites, etc ....

Future is very close....FEARING !!!

Sunil

Permalink Reply by Ritesh Kumar Tiwari on April 26, 2012 at 19:16

Dear Sir,

Thanks for such a brief answer. Sir, but i think it would take a long time, because parallel to other subjects i.e., chemical engineering, Pharmacy, Medicine, some Intelligent systems have been developed in LIS also. But a study says that, Intelligent System developers of same have found the classification as most challenging aspect in comparison to other tasks of Library Management.

I have found a article based on the study of AI and IS for  LIS. Please find the same as attachment. Please refer the same it's a good article. 

After reading the article I found that, only fully faceted classification scheme can support the AI (as Dr. Ranganthan has always emphasized on Fully Faceted Classification Scheme). otherwise  it is very difficult to develop an IS for Classification with rigid & annumerated classification scheme.

It's my vision but I want to know other's.

Attachments:
Permalink Reply by Alka Gohel on April 26, 2012 at 19:27

I want to know more on this topic

Kindly inform

Permalink Reply by S muansangluai Ngaihte on April 29, 2012 at 6:33

yes of course with the growth of Inforamtion technology alots of software come up which are compatable for mangement and administration of documents available in the library, such that softcopy some classification documents such as DDC are available too in cd format which are considered to be more convinient for classifying documents or books....apart form that UDC too have a soft copy on that...

Permalink Reply by Siddhartha Shankar Ray on April 29, 2012 at 21:43

Dear Sri Ritesh Kumar Tiwari,

Thanks for your post, As I feel theoretically it is possible provided you should have an online context vocabulary. However, the orientations of interdisciplinary materials seems to be complicated. For example, I worked in a Medical Library and books on Social & Preventive Medicine were classified under 610 as Main Class but here at my present place which is a Social science Library, the same books (used for teaching & research of Medical Sociology and Health Economics) are Classified under 310 or 330 as main class. Another complication is that the automated Class numbers should presumably be mechanically based upon key words and text words. The Classifier have to ascertain both syntactic and semantic relationship among these text or key words. Those who has done some in-depth studies with PRE CO-ORDINATE INDEXING specially PRECIS, especially with its role operators may support my views.At this moment, I have with me a book entitled International Relations Theory and South Asia; Vol. 2; Security, Political Economy, Domestic Politics, Identities and Images edited by E. Sridharan, New Delhi, Oxford University Press 2011; ISBN: 0198070802. with all these text words and with the intrinsic key words, it seems to be really complicated to classify with automated classification schemes. Lastly, admittedly, the confusion will be less in enumarative classifications but will be very high fot analytico-synthetic schemes of Classifications,

However, Sri Ajit Prabhakaranhas suggested about I-Google, and specified its effectiveness in converting conTEXTS to conTENTS. I have not tried it so can not say about its feasibility, but if it really happens as Sri Prabhakaran suggests, it may be possible with requisite modifications.

Sincerely,

Siddhartha S. Ray

Permalink Reply by Ritesh Kumar Tiwari on April 30, 2012 at 20:19
Thanks for your kind words !
Permalink Reply by Ritesh Kumar Tiwari on April 30, 2012 at 20:21
Dear Alka madam,

Thanks for showing your interest. I shall try my level best to provide you more material on this topic. Kindly give your email address.
Permalink Reply by Ritesh Kumar Tiwari on April 30, 2012 at 20:25
Dear Mr. Ngaihte

Thanks for your kind reply.

CD Format of DDC has its own complications. If you had used it you could understand. It takes too much time to search the keywords. It's not the example of Artificial Intelligence at all.
Permalink Reply by Ritesh Kumar Tiwari on April 30, 2012 at 20:31
Dear Mr. Ray,

Thanks for your kind reply !

I do agree with you at all points. The most typical part of the Intelligent system is to make it able to understand the subject by reading the scanned text. It seems very difficult because so many times the intelligent systems (Ex. made for translation) make errors in reading the text itself, than image how difficult it would be to understand the text for an Intelligent System.

Thanks again for your kind reply and brief explanation.

RSS

© 2024   Created by Dr. Badan Barman.   Powered by

Badges  |  Report an Issue  |  Terms of Service

Koha Workshop