The invention of Internet, sensors networks, and IoT facilitated real-time data extraction as precisely as possible, adding features to the data volume, velocity, and variety. The technological growth in the storage systems and high-performance computing gave the opportunity to design the computational strategies to calculate required statistics and then information processing for making decisions to address the real-world problems preferably just in time. The computational methods should sink with various natural topological configurations of data management systems, communication systems, which avoid frequent visits (load, unload) of voluminous raw data. The effectiveness of a recommender system is influenced by all these above factors. One need to design hierarchical and hybrid approaches for building essential required information representation/processing for decision-making process at higher level of abstraction. One of the possible approaches can be designed by exploiting distributed computing paradigm.