Survey: A Hybrid Approach to Solve Cold-Start Problem in Online Recommendation System
27 Pages Posted: 9 Mar 2018 Last revised: 18 Apr 2018
Date Written: November 15, 2017
Abstract
The online recommendation system is prevalent in social networks, the primary goal of online recommendation system is to recommend the best suitable items to the user. In the recent year online recommendation systems do the leading role in filtering and to make something to a customer specification to the information. There are various recommendation techniques which are used to recommend a product to the active user appropriately. However, it’s hard to recommend items what the point when purchase history of client is not available. This issue is known as user cold-start problem. Although many efforts have been made to solve cold start problem, it is still a challenging issue in commercial sites. In this paper is an excellent survey focusing on issues of cold start recommendation. Now the biggest problem is sometimes when user come what too recommended. The other issue is regarding providing the item in proper order otherwise wrong item may misleading to other valuable items. One more important issue is also regarding the data available for analysis. This data may contain lots of Noisy data and if they start analysis using this type of data may end in the wrong or unwanted result.
Keywords: Online recommendation, Cold-start, Collaborative filtering, Content based filtering, Hybrid filtering, Social networks
Suggested Citation: Suggested Citation