Towards Scalable Recommendation Framework with Heterogeneous Data Sources: Preliminary Results
Nam D. Vo, Jason J. Jung · 2018
This paper presents the basic concepts of a scalable recommendation framework (called DakGalBi) that can integrate all possible heterogeneous data sources to provide users with optimized recommendations. The framework consists of three components: the database management system (HBase), machine-learning engine (Spark), and indexing module (Elasticsearch). Hence, the framework enables recommendation systems to deal with heterogeneous data sources in various recommendation scenarios. Our early implementation proved that DakGalBi has significant performance in terms of working with heterogeneous data sources.