Study Of A Library / Catalog Recommender System


Summary

Understanding the functionality and scope of betterment of a library / catalog recommender system BibTip http://www.bibtip.com/en


Catalogue Recommender System : BibTip

BibTip is a recommender system that adds value for online catalogues.

BibTip recommends links of similar titles through careful observation of user behaviour during catalogue search and statistical analysis

Data generated via title displays in online catalog sessions are stored and user views counted. The co-occurrences between pairs of titles are established if the two titles were viewed together in at least one session. These co-occurrences are counted and summarized in a co-occurrence matrix. In a further step, this matrix is evaluated in order to generate the recommendations

It uses Repeat Buying Theory.


Scenario Design & Analysis:

Target Users And Goals

Target Users of BibTip are the the online library users.

The key goals of this product is to study user preferences and generate custom-tailored catalog-specific recommendations.

The goals can be met by observing and aggregating user behaviour and browsing history, wishlists, to-read lists, and making recommendation based on these and matching with read-lists of other users with similar interests.


Reverse Engineering:

The product can maintain information (to-read lists, read-lists of other users and make it available for views (depending on privacy settings opted by the user)

Also depending upon user profile and background (industry/stream the user belongs to) recommendations model can be made and applied to new library user sign-ups.


Recommendations For Product Improvement

Cross Library linking and recommendations can be done by integrating different college libraries from same university / state / public libraries

Advanced administrative and customizing tools for libraries can be provided

Use clustering techniques to to facilitate advance subject wise searches