Decoding plastic recycling in Vietnam: Packaging design, consumer recognition and model-based exploration

This study investigates plastic food packaging (PFP) recycling symbols in Vietnam through field surveys, questionnaires and statistical and machine-learning models. Results show that 68.2% of shoppers correctly identified the recycling symbol, whereas 87.2% confused waste disposal symbols with recyclability. Age correlated positively with accuracy (<i>p</i> < 0.01), but gender had no effect. Waste separators performed worse, likely due to misconceptions (<i>p</i> > 0.05). XGBoost model achieved the best prediction accuracy (MacroF1 = 77.6%, kappa = 0.33), followed by Random Forest (MacroF1 = 76.0%, kappa = 0.29). Shopping frequency, perceived importance of recycling and willingness to support recyclable products were stronger predictors than demographics. Analysis of PFP revealed that 66.7% was recyclable and 33.3% non-recyclable, with polyethylene terephthalate (46.5%) and polypropylene (40.1%) being the dominant recyclable types. Findings highlight the need for standardized labelling, mono-material designs and predictive modelling to enhance consumer awareness and improve recycling efficiency.