• It is a facility that involves predicting user responses to options in web applications.

• For example web search recommendation, product recommendation, friend recommendation in social media, etc.

• There are two types of recommendation systems:

a) Collaborative filtering.

b) Content based filtering.

Collaborative filtering systems:

• It uses community data from peer groups for recommendation.

• These exhibits all those things that are popular among the peers.

• These filtering systems recommend items based on similarity measure between users and/or items.

• Here user profile and contextual parameters along with the community data are used by the recommender systems to personalize the recommendation list.

• This is the most prominent approach in e-commerce site.

Example:

• Consider a movie rating system.

The basic assumption for collaborative filtering is :

• User gives ratings to item in the catalog.

• Customer who had similar taste in past will have similar taste in future.

• Users who agreed in their subjective evaluations in the past will agree in the future too.

• To find out similarity we can use Pearson's correlation co-efficient as:

• $sim (a,b) = \sum_pep \frac{(ra, p - ra) (rb, p - rb)}{\sqrt{\sum_pep (ra, p - ra)^2, (rb , p - rb)^2}}$

Where A, b = users

Ra, p – rating of user ‘a’ for item ‘p’

P – set of items rated by both a and b

• We can use this formula for prediction as :

$pred (a,p) = ra + \frac {\sum_ben sim (a,b) \times (\eta b , p - r b)}{\sum_ben sim (a, b)}$

• We can apply this to movie rating system and based on that we can predict rating for movie as well as to whom the movie should be recommended.