Household cooking in the context of carbon neutrality: A machine-learning-based review

Abstract

About 6.7% of global greenhouse gas emissions is caused by household cooking activities and thus it is of significance to identify research gaps between current studies and future directions in the context of carbon neutrality. To this end, the Latent Dirichlet Allocation topic model is used to review a total of 1440 household cooking studies from international journals written in English between 1983 and 2021. The textual mining technique helps to identify 20 topics in machine-learning sense, involving 8 research disciplines. In addition to energy field, household cooking is most relevant to disciplines of Multidisciplinary, Clinical Medicine, Chemistry, Economics and Business, and Geosciences. Energy ladder hypothesis and energy poverty are the most prevalent topics and asymmetric dependence relationships are unveiled among the 20 topics. Almost all cooking topics focus on health risk elimination and the transition to cleaner fuels while the target of carbon neutrality has not been adequately considered. The practical cooking fuel transition pathway, health co-benefits, impacts of the shift in cooking methods and practice on cultural diversity and human society driven by carbon neutrality constitute potential research directions. The machine-learning literature review research framework used in the study can be generalized in era of big data.