The Impact of Randomized Algorithm over Recommender System

• Published in: Procedia computer science 2021, Vol.194, p.218-223
• Main Authors: Shakil, Ubaid, Syed, Alam, Mohammed Talha, Sohail, Shahab Saquib, Saifi, Imran Khan, Mufti, Tabish, Aziz, Asfia, Nafis, Md Tabrez
• Format: Article
• Language: English
• Subjects: Collaborative Filtering; Machine learning; Randomization; Recommender systems
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2021.10.076

In today’s world, the amount of information is increasing drastically. There are 2.5 quintillion bytes of data generated each day at a current time. With this rapid increase of data, it is important to filter the huge amount of data from the existing unnecessary available storage. To handle this issue, the researchers have incorporated the concept of recommender system. Recommender system tries to filter the users’ data to come up with the exact amount of information required according to the needs and preferences of the users. Hence there is an increase in the demand of recommender system to tackle the data overload. However, the existing recommender systems do perform well but face some issues in the corporate world. Cold start, user bias, precision and speed are some of the major problems faced by recommender system which are yet to be addressed. To minimize such issues, we incorporate the concept of randomization such that random factor can be integrated in design of recommender systems. In this paper, we have also outlined the advantages as well as the procedure of using a randomized algorithm for recommender system. The proposed approach would provide a platform for researchers to enhance the recommender system using the randomized algorithms in the future.