Innovative technological modes of Data Mining and modelling for adaptive project management of food industry competitive enterprises in crisis conditions

Abstract

Developed in this research scientific and practical applied project solutions regarding Data Mining for enterprises and companies (on the example of food industry) involve the application of advanced cybernetic computing methods/algorithms, technological modes and scenarios (for integration, pre-processing, machine learning, testing and in-depth comprehensive interpretation of the results) of analysis and analytics of large structured and semi-structured data sets for training high-quality descriptive, predictive and even prescriptive models. The proposed by authors multimode adaptive Data Mining synergistically combines in parallel and sequential scenarios: – methods of preliminary EDA, – statistical analysis methods, – business intelligence methods, – classical machine learning algorithms and architectures, – advanced methods of testing and verification of the obtained results, – methods of interdisciplinary empirical expert interpretation of results, – knowledge engineering formats/techniques – for discovery/detection previously unknown, hidden and potentially useful patterns, relationships and trends (for innovative project management). The main methodological and technological goal of this developed methodology of multi-mode adaptive Data Mining for food industry enterprises is to increase the completeness (support) and accuracy of business and technical-technological modeling on all levels of project management of food industry enterprises: strategic, tactical and operational. By optimally configuring hyperparameters, parameters, algorithms/methods and architecture of multi-target and multidimensional explicit and implicit descriptive and predicative models, using high-performance hybrid parallel soft computing for machine learning – the improved methodology of multimode Data Mining (proposed by the authors) allows to find/detect/mine for new, useful, hidden corporate knowledge from previously collected, extracted, integrated Data Lakes, stimulating the overall efficiency, sustainability, and therefore competitiveness, of food industry enterprises at various organizational scales (from individual, craft productions to integrated international holdings) and in various food product groups and niches.