For predicting the correct choices and improving user experience of a product, recommendation systems are essential in research to prove their productivity. The goal of the scientometrics study is to recognize growth of the Recommendation system by implementing deep learning algorithm. This paper shows the research work in deep learning-based recommendation system through bibliometric study by exploring the twenty years of work from 2001 to 2020. Scopus is the largest database which consists of information about abstract with citation databases of peer-reviewed in various areas. study gathers the publication from Scopus database and retrieves total 6,813 publications in different types such as conference paper, journal and reviews articles, book chapter etc. the China and India are the most prominent countries for publishing recommendation system research work. Keywords plays the important role to search the documents, in this study collaborative, content and deep learning are the important terms to identify the documents. The computer science, decision science, mathematics and Engineering are effective research subject areas were recommendation system work, Hence the deep learning base recommendation system applicable to all domain.