Recommender Systems (in brief)

One of the artificial intelligence technologies that in relevance to LIS discipline is ‘Recommender System’ which is ‘An information filtering technology.’ usually through using different methods as following:

*Content-based systems which are ‘systems recommending items similar to items a user liked in the past'(Krasnoshchok, 2013), in other words, it recommends items by its features.

According to Krasnoshchok the advantages to using this method are:
-It can provide a recommendation as soon as it has information of items available.
– It does not need any user data to provide a recommendation.

On the other hand, the disadvantages can be that:
-Items and attributes must be machine-recognisable.
-The absence of personal recommendations.
-No serendipitous items.

* Collaborative filtering systems which ‘recommend items based on similarity measures between users and/or items'(Leskovec, Rajaraman, and Ullman, JD, 2014). It can be user-based collaborative filtering or item-based collaborative filtering.

According to Krasnoshchok the advantages to using this method are:
– CF methods utilise only ratings and do not require any additional information about users or items.
-These systems can make an assessment of quality, style or viewpoint by consideration of other people’s experience.
-CF systems can produce personalised recommendations because they consider other people’s experience and recommendations are based on that experience.
-CF recommender systems can suggest serendipitous items by observing similar-minded people’s behaviour.

While the disadvantages are:
– The system cannot provide recommendations if ratings weren’t available.
– The accuracy of the recommendations would be poor when there is few data about users’ ratings. Which this defined as ‘Cold-Start problem’.
– Many of existing CF algorithms work slowly on a huge amount of data, which was the reason to create several techniques such as clustering and parallelization to overcome the problem.

There are also more methods of recommender systems namely: personalised learning to rank, demographic, social recommendation, and hybrid.

Examples of Recommender Systems Libary:

  • Mendeley
  • BookPsychic
  • LibRec
  • MyMediaLite

Finally, this was a brief information of recommender systems which is a way to help us saves time by discovering items that we may not find easily.

Reference:

*Krasnoshchok, O. (2013) Content-based Filtering Recommender systems – benefits and disadvantages – Recommender systems / recommendation engines explained. Available at: http://recommender.no/info/content-based-filtering-recommender-systems/ (Accessed: 4 December 2016).
*Krasnoshchok, O. (2013) Collaborative Filtering Recommender systems – benefits and disadvantages – Recommender systems / recommendation engines explained. Available at: http://recommender.no/info/collaborative-filtering-approach/ (Accessed: 4 December 2016).
*Leskovec, J., Rajaraman, A., Labs, M. and Ullman, J.D. (2016) Mining of massive Datasets. Available at: http://infolab.stanford.edu/~ullman/mmds/book.pdf (Accessed: 4 December 2016).
picture:
http://www.isoin.es/en/blog-en/
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s