Requirements Engineering in Machine Learning Projects

Over the last decade, machine learning methods have revolutionized a large number of domains and provided solutions to many problems that people could hardly solve in past. The availability amounts data, powerful processing architectures, easy-to-use software frameworks made popular, readily available, affordable option different contexts. However, development maintenance production-level systems proven be quite challenging, as these activities require an engineering approach solid best practices. Software offers mature process practices for conventional systems, but some them are not directly applicable new programming paradigm imposed by learning. same applies requirements Therefore, this article provides overview challenges been reported research literature, along with their proposed solutions. Furthermore, it presents our overcoming those form case study. Through mixed-method study, tries identify necessary adjustments (1) (2) understanding certain types better fit specifics Moreover, emphasize relevance properly conducted addressing complexity well motivate further discussion on developing such systems.