Abstract
The agrarian economy, with its social agents and technical systems, mobilizes the elements that generate transformations in the social and natural landscapes in the Brazilian Amazon. Choices for regional development lead to sustainability or unsustainability of the forest ecosystem and its social landscape, while not including health in this debate. We argue that an analytical framework for integrated health-environment-economy approaches needs a territorial representation for the landscapes associated with the ways of living and producing in Amazonian agriculture: the production landscape units (PLU). In this article, we explore machine learning techniques, in the field of supervised classification, with methods based on decision trees, to identify and map the PLU. A case study is developed for the municipalities of Mocajuba and Cametá, in the Baixo Tocantins region, in the State of Pará, for 2021. We describe how to identify and map the PLU in an intra-municipal spatial unit of reference and how to associate them with the types of rural techno-productive trajectories or technological trajectories (TTs) found in the regional agrarian economy. We promote an initial discussion on the use of PLU in the structuring of integrated approaches in health. This article contributes to align debates on strategies for economic development with health promotion in the Brazilian Amazon.