Addressing Cold Start Problem in Recommendation System Using Custom Built Hadoop Ecosystem

摘要

In the modern digital era, the most common problem is being surrounded by the e-Commerce and social networking sites which provides accurate and timely recommendations based on user interests. When we have user-related data, it will be easier to give recommendations to those users based on their previous activity in that particular website. But cold start problem arises when there is no information related to a user or an item, because they are new to that particular website. Hadoop is one of the distributed huge data processing framework that provides an effective solution to overcome these cold start problems in real time scenarios. Here, we have used the tools like Sqoop, Hive and build a model that address and resolve certain cold start issues in recommendation systems.