Tomislav Hlupić works as a consultant in Poslovna Inteligencija, a regional leader in BI solutions & implementations, where he’s gaining expertise in banking and retail BI systems, working in Microsoft BI stack. Apart from that, he is on his 2nd year of PhD studies with the focus on structured and unstructured data integration, real-time analysis and Data Lakes. Through the research he combined his passion for advanced analysis technologies and systems and IoT, trying to augment the visible data with hidden data all over the environment. Together with colleagues from his company, he is involved in conducting several subjects at Algebra University College on the graduate level.
Building a recommendation system for IPTV on a fast streaming architecture
The talk will be a mixture between a description of fast streaming architecture on which the system was built on and the IPTV recommender system that Poslovna Inteligencija has developed. The overview of the topic will be given in the introduction, following by a description of content delivery services and the data produced by them and how it is used in the customer experience. Next, the overview of recommender system will be given together with the architecture of the content analytics system and the implementation (including the algorithm) of the recommendation engine, which is a part of the content analytics system.
We will cover the architecture of the system and the implementation equally, not giving the priority to any of the parts. We plan to present the reasons for the system creation, the progress in the architecture based on tests and the final implementation of the algorithm, together with the presented difference between various algorithms used for same or similar purposes. The content of the topic is backed with the article on the mipro 2018 conference and is the current highlight of the Innovation & Development department in Poslovna inteligencija.