Figure 10
The framework places model quality at the centre, connected to data-centric, algorithmic and methodological, complexity, and computational and infrastructure dimensions. Data-centric factors include data quality dimensions, data quantity and distribution, and stakeholder participation and data contribution. Algorithmic and methodological factors include data processing and engineering, algorithm design, and training and testing strategies. Complexity includes problem scale, influencing factors, objective functions, dynamism, variable relationships, temporal-spatial dependencies and uncertainties. Computational and infrastructure factors include computing resources, processing power, scalability, memory, power consumption, software and hardware dependencies, and operational trade-offs. Environmental and domain-specific factors include the operational environment, disaster characteristics, humanitarian context and supply chain context. Arrows connect these factors and indicate relationships among data requirements, algorithm choices, training requirements, computational needs and model quality.Summarised factors affecting HSCML model quality from the literature
Source: Authors’ own work