A Study on E-commerce Recommender System Based on Big Data
Abstract : Recommender system algorithms are widely used in e-commerce to provide personalized and more accurate recommendations to online users and enhance the sales and user stickiness of e-commerce. This paper discusses several recommendation algorithms and the challenge of tradition recommender system in big data situation, and then proposes a framework of distributed and scalable recommender system based on Hadoop. The recommender system based on Hadoop, combining the advantage of computational ability and scalability of Map Reduce and hybrid recommendation algorithms, brings a solution to information overload problem in big e-commerce .
E-commerce recommender systems have been converted to a very important decision-making helper for customers, and provide online personalised recommendations using information technology and customers’ information.
The recommender system is a technology that helps to create personalisation features.
A satisfactory and new area in the field of the RSs is to analyse the characteristics and potential interests of customers (Salehi, 2013; Sharma and Ray, 2016).
In EC, RS evaluate the purchasing behaviours of the customers and learn from these behaviours to try to recommend suitable goods
The main disadvantage of Content-based Filtering is that its effectiveness greatly depends on the extensive and in-depth description of products’ attributes.. Another problem is that Content-based Filtering only recommend items similar to the products described in users’ profile with the problem of content over-specialization. With the combined advantages of Hadoop distributed computation ability and mixed recommendation, the scalable, flexible and diversified recommender system can obviously bring a solution to information overload problem in big e-commerce and provide sustainable competitive advantage for e-commerce with personalized marketing.
The proposed system of this research has resulted in the removal of traditional CF constraints and presents more accurate and appropriate recommendations for the preferences of customers.
This article provides an overview of the current state of the art in recommendation systems, their types, challenges, limitations, and business adoptions.
To assess the quality of a recommendation system, qualitative evaluation metrics are discussed in the paper.
Various application shave adopted recommendation systems, including e-commerce, healthcare, transportation, agriculture, and media. This paper provides the current landscape of recommender systems research and identifies directions in the field in various applications.
? Collaborative Filtering is widely used in e-commerce and is one of the most successful recommender systems in e-commerce.
? Collaborative Filtering focus on finding the nearest neighbors of target user who either purchased the same precuts or rated similarly on the same products with target user .
? The commonly used algorithm is Content-based Filtering is Association Rule Mining, which is a very useful method to discover the relationship hidden in large datasets.
? This layer includes several data mining algorithms and machine learning algorithms, such as Association Rules mining, Clustering and Collaborative Filtering.
? The algorithms layer is independent from recommendation layer for the purpose of integrating more algorithms.
|