Tuesday, March 4, 2014

Intelligent library systems and big data


What do I mean by Intelligent Library System? To my view an intelligent library system is kind of a loosely coupled system having an in built intelligence to serve each user based on his/her need proactively and track each of them based on the areas below:

  • Work Profile - This includes the work status like student, researcher, working or simply a learner
  • Personal Profile - Includes his social status, likes and dislikes, friends and families
  • Primary areas of interest - The primary or core areas of interest
  • Secondary areas of interest - Any secondary areas he/she is interested on
All these information defines the meta data of a user and this actually gets linked to the information access history of that user over a period of time. Though the access pattern may seem random over a short window but there should be a definite pattern of information access over a bigger period of time. The user metadata along with the pattern can lead us to interpolate future information need for that user and that would be the intelligence of the system. Now the metadata I talked about can be static or dynamic in nature which basically gets modified by itself based on the user activity in other linked sites (mainly the different social networking sites) and his/her information access pattern. The core intelligence engine would take out this information and analyze to determine the state of the user's information need. The state of a user's information need defines probable information requirement of that user in different time scale say within a week or in a month and so on. So based on the state and the information availability across the library the system can get the data either from the library or from external sources, cache that information and tag it with the user profile. This is kind of getting prepared with the information early on the bed so that user can instantly get the information. This is kind of an illusion to the user who believes the service has just been in real time.    

So ideally, the library system would behave as many parallel virtual libraries for each of it's users and each of the virtual instance would be responsible for working out the respective intelligence catered for that user. Now if a library has a sizable number of users and grows continuously then we can feel the huge amount of data sets it would be building over a period of time. And since, there is a constant need to access this huge amount of information in order to intelligently decide the user's need, the retrieval has to be really fast. Also, keep in mind the metadata information and the search pattern could include various types of data like text, pictures, movies, blobs, podcast and so on. Hence we are more closely dealing with the concept of big data that actually provides ease of access to a wide variety of vast amount of data in real time. 

So an intelligent library system is more likely to use the big data concept in order to decide the user's need in real time and provide a more meaningful information that the user deserves.