A Recommender System for Machine Learning Pipelines' Design to Recognize Objects in Video Frames as a Learning Tool for Training Data Scientists

This work-in-progress describes an approach to use a recommender system for machine learning pipelines' design to recognize objects in video frames as an electronic mean to train data scientists. We use the object-oriented entity called “Abstract hybrid learning resource” (AHLR) to unite in one information structure learning objects, data resources, text resources, application software, computing libraries, and instructional descriptions. The polymorphism features AHLR allows us to make various options of learning resources according to learning conditions, academic group and stunt's individual parameters, type of learning activities, etc., based on the eponymous method like “Lecture”, “Project, etc. The paper shows the use of a recommender system and its products, including recommended/generated pipelines, machine learning libraries, manuals, tutorials, and related scientific papers an information resources to build needed learning resource.