Recommendation

Movie Recommendation System

A. Content Based :

Recommend Product or Content based on Similar Attributes.

• I Saw a Comedy Movie Hera Pheri and Rated 5 Star
• I Saw a Comedy Movie Dhamaal and Rated 5 Star
• YouTube Started Recommending me more Comedy Movies ( Content )
• Advantage : Works even when Product or Content has no Rating but Attribute is Important for every Product or Content.

B. Collaborative Filtering :

Recommend Product and Content based on Past User Ratings not based on Attributes.

• I Saw Avengers and Rated 5 Star | My Friend Akash Saw Avengers and Rated 4 Star
• I Saw Underground and Rated 4 Star | Akash Saw Underground and Rated 5 Star
• I Saw Captain America and Rated 5 Star | Hotstar Recommended Captain America to Akash.
• Advantage : No need of Attributes, but User Reviews are very Important.

Real Life Examples

1. Social Media

• Facebook : Posts and Friends.
• Pinterest : Products, Posts and Pins.
• Instagram : Posts, Timeline and Friends.
• Online Content : YouTube, Hotstar and Netflix ( Shows, Movies and Contents )

1. Music Service : YouTube Music, Spotify ( Playlist ) and Apple Music Songs.

2. Ecommerce : Amazon, Flipkart and Myntra Products.

3. Banking and Insurance : Best Policy, Loan Amount and Insurance.

Recommender System ( Takes User Past Experience )

• Predicts the Rating the User would give the Product.
• Recommend the Products which will be liked by the User.
• Rank Products by User Preference.
• Product Similarity ( If User Buy Mobile he will also Buy Case )
• Find Similar User Preferences. ( Peoples mostly Buy this also )

New User

• If the User is New, He / She cannot be Recommended unless any Rating is Submitted.
• Until then Products Suggested are Based on his Search Interests and Similar Products.

Handling First Time Users

• Do not make any Recommendations.
• Recommend based on Product Similarity.
• Recommend based on overall Average Ratings.