Ente Varthakal

Inspiration

We were always passionate about implementing new ideas in Artificial Intelligence and Machine Learning.We thought about how we can use this techniques in Media industry and came up with this idea.

What it does

It studies the behaviour of the online news readers hence recognising their reading patterns and interests.This information collected will help suggest most personalised content which users were always interested in.

How we built it

We use Mahout for the recommendation.Front End is developed using HTML5, CSS3, Angular JS.The Back End development is carried out using Java with Spring and Hibernate.The database used to save user data is MySql from which data is served to Mahout which in turn provides the recommendation data. The semantic search data is mapped and saved using Neo4j and the API for seving this information to the front end is done using Flask.

Challenges we ran into

Getting a recommendation pattern for a user was a big challenge. Implementing Collaborative Filtering by combining User based recommendation and item based recommendation was a bit of a task altogether. Implementing the Personalised Semantic search algorithm was challenging.

Accomplishments that we're proud of

Implementing the Collaborative filtering (User based recommendation and item based recommendation) in such a short span. Saving the Mapped data in Neo4j. Building the API to implement Personalized Semantic search.

What we learned

To complete a full fledged web app with user recommendation and semantic search is not a task for a day but everything has its starting point. Working in a hackathon environment is a challenging thing.