Despite the hype around data analytics, the success rate of analytics initiatives remains very low and the value of data in organisations is left hidden. Various research studies show that the main barriers to analytics adoption are organisational and the lack of structured approaches on how to conduct analytics initiatives is a possible cause of analytics project failures. Data mining process models then become fundamental means to support analytics project management and minimise the risk of data dredging. In this paper, Knowledge Discovery and Data Mining process models are reviewed starting from the most popular models currently in use. Four distinctive research paths for data mining process models have emerged. These evolution paths seem to address limitations of the CRISP-DM model which remains the de facto standard in industry. The research streams identified include the evolution of the human role; the relevance of iteration and interactions; the role of data and knowledge repositories; and the integration of software engineering/agile methodologies. In the future, these four research streams should be combined to support the development of more encompassing process models.
The implementation of Discrete Event Simulation (DES) -based decision support tools in complex manufacturing environments could prove of invaluable help to industrial practitioners involved in cross-functional decision processes at multiple hierarchical levels. The increasing number of decision variables, their stochastic nature and the non-linearity of their mutual relationships theoretically make simulation a preferred modelling approach for a great variety of manufacturing systems as strict simplifying assumptions are not necessarily required and the models' detail level can be tuned according to the analysis purposes. However, recourse to Commercial Off-The-Shelf (COTS) simulation packages to develop and implement simulation-based solutions in real manufacturing environments usually presents significant cost-of-ownership (COO). Along with license costs, modelling flexibility and sustainability represent fundamental issues raised by industrial engineers that adopt COTS simulation packages. In order to promote the use of DES in production related decision making processes and reduce the associated COO for manufacturing companies, an open-source simulation platform, ManPy, has been developed. ManPy consists of a library of DES objects implemented in SimPy. ManPy's scope is to provide modellers with generic, highly customizable open-source simulation objects that can be connected to form a model in the same fashion of COTS simulation packages. ManPy's on-going development is based on guidelines provided by the analysis of real industrial use cases. Specific pilot models developed in SimPy are used to identify new objects and relevant features to be incorporated in ManPy in order to make it a highly flexible simulation tool. In this article, a use case based on a labour intensive serial production line operating in a medical device manufacturing plant is described. Insights for the transition from a COTS simulation model to a specific SimPy model and finally to generic ManPy objects are presented.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.