Using the Machine Learning Method to Study the Environmental Footprints Embodied in Chinese Diet.

Area of Science:

• Environmental Science
• Agricultural Economics
• Data Science

Background:

• The food system significantly impacts environmental sustainability and resource use.
• Chinese residents' food consumption patterns exert considerable environmental pressure.
• Individual differences in food consumption necessitate targeted approaches to mitigate environmental impact.

Purpose of the Study:

• To develop a machine learning-based method for identifying populations with high environmental footprints.
• To analyze the relationship between dietary intake and environmental resource consumption in China.
• To pinpoint demographic characteristics influencing high consumption tendencies.

Main Methods:

• Utilized microdata from the China Health and Nutrition Survey (CHNS).
• Employed a decision tree algorithm for user identification and characteristic analysis.
• Investigated the link between dietary patterns and environmental footprint.

Main Results:

• The environmental impact of the food system follows a logistic normal distribution trend.
• Gender, income, education level, and region were identified as the most influential demographic factors on environmental footprint.
• The decision tree model effectively identified high-consumption populations with improved accuracy and coverage.

Conclusions:

• Machine learning, specifically decision trees, can accurately identify individuals with high environmental footprints.
• Understanding demographic drivers of consumption is crucial for promoting sustainable food consumption patterns.
• Targeted interventions based on identified characteristics can enhance environmental sustainability in China's food system.