Modelling ragpickers’ productivity at the bottom of the pyramid: the use of artificial neural networks (ANNs)
本研究运用人工神经网络分析印度拾荒者生产力的非线性影响因素，发现对非政府组织的接受度、识字率、设备技术投入和团队规模是关键变量，为提升拾荒者收入和回收率提供路径。
Purpose In this research, the authors apply artificial neural networks (ANNs) to uncover non-linear relationships among factors that influence the productivity of ragpickers in the Indian context. Design/methodology/approach A broad long-term action research program provides a means to shape the research question and posit relevant factors, whereas ANNs capture the true underlying non-linear relationships. ANN models the relationships between four independent variables and three forms of waste value chains without assuming any distributional forms. The authors apply bootstrapping in conjunction with ANNs. Findings The authors identify four elements that influence ragpickers’ productivity: receptiveness to non-governmental organizations, literacy, the deployment of proper equipment/technology and group size. Research limitations/implications This study provides a unique way to analyze bottom of the pyramid (BoP) operations via ANNs. Social implications This study provides a road map to help ragpickers in India raise incomes while simultaneously improving recycling rates. Originality/value This research is grounded in the stakeholder resource-based view and the network–individual–resource model. It generalizes these theories to the informal waste value chain at BoP communities.