Streamlined News Recommendation System Using a Variable Markov Model

Providing news to users in a news article recommendation system is a balancing act between delivering news that is recent and news that is relevant to their interests. Users should be able to receive a stream of similar articles that interest them and control their traversal through the topics of news articles in a stream-wise fashion as well. A Variable Markov Model (VMM), built on trends in recently published news articles, is proposed as a single solution to categorically cater news to all users with minimal overhead and maintenance. This single model provided to all users throughout experimentation has shown that, though it is not built based on user interests, it is applicable as a basis for applying user interest and trend factors upon to achieve catered and novel news recommendation experiences.