Journal of Consumer Protection and Food Safety最新文献

Machine learning has been extensively used for analyzing spectral data in food quality management. However, collecting high-quality spectral data from miniature spectrometers outside the laboratory is challenging due to various factors such as distortions, noise, high dimensionality, and collinearity. This paper presents an in-depth analysis of food datasets collected from miniature spectrometers to evaluate the data quality and characteristics, by focusing on a case study of olive oil quality check, where various machine learning models were applied to differentiate pure and adulterated olive oil. Furthermore, the impact of pre-processing techniques on data distortions was studied. It presents a comprehensive pipeline, including data pre-processing, dimension reduction, classification, and regression analysis, and deploys different algorithms for comparative classification and regression analysis. The model performances were assessed using 2 separate methods: tenfold cross-validation on an entire dataset with 10% random testing, and an entire test set collected in different environments (multi-session validation). The first validation approach reached classification rates of up to 96.73%, while the second achieved 83.32%. These results demonstrate that cost-effective miniature spectrometers augmented with a suitable machine learning pipeline could execute classification tasks on par with non-portable and more expensive spectrometers. Furthermore, the study highlights the requirement of specialized algorithms to handle different ambient conditions affecting data acquisition and to eliminate performance gaps, making miniature spectrometers suitable for in situ scenarios. This work extends previous research to enable consumers becoming the first line in the defense against food fraud.

Zusammenfassung

The food industry has been greatly impacted by COVID-19, causing governments to restrict food exports to prevent shortages. A negative food trade balance reveals a country's dependence on imports and underscores the significance of a sound food policy. Hence, for the first time, this study examines the J-curve hypothesis for the U.S. with Canada at the state rather than country level and creates maps based on the findings. The approach of this study differs from all empirical studies using country-level J-curve analyses, because the U.S. may require a state level analysis since its states differ in terms of economic-population sizes, tax rates, and administrative structures. For this aim, this study employs the linear and nonlinear autoregressive distributed lag (ARDL) approaches. The results indicate that while only 8 out of 47 U.S. states support the food-based asymmetric J-curve hypothesis, 15 U.S. states support the asymmetric inverse J-curve hypothesis. Additionally, 9 U.S. states support the food-based symmetric J-curve hypothesis, and 2 U.S. states support the symmetric inverse J-curve hypothesis. Based on these results, policymakers of U.S. states where the J-curve hypothesis is not supported should review their food-based bilateral trade policies with Canada.

Graphical abstract

These maps depict the U.S. states in green and red, indicating support for the J-curve and inverse J-curve hypotheses, respectively. The map on the left was generated using the linear model (symmetric approach), while the map on the right was generated using the nonlinear model (asymmetric approach).

The German National Reference Centre for Authentic Food (NRZ-Authent) and the competent German food control authorities of the federal states cooperated within the framework of the 10th joint Europol INTERPOL operation OPSON (OPSON X) in the detection of adulterated meat products. A total of 63 meat product samples were collected and analysed by the authorities using standard analytical procedures and subjected to a recently published 16S rDNA metabarcoding analysis. The sequence reads were analysed using 3 bioinformatics data processing strategies. The study aimed to gain additional data on the test samples regarding the authenticity of the declared species and to validate the 16S rDNA metabarcoding method with representative samples. The method was tested not only on 63 test samples, but also on 5 commercial samples from 2 interlaboratory comparison studies and 9 mock mixtures in parallel. The 16S rDNA metabarcoding method was able to detect species that were not target species of the used standard analytical methods, but failed, as shown previously, to detect fallow deer. Otherwise, the qualitative results of the 16S rDNA metabarcoding method were very similar to those of the methods currently in use by the German food control laboratories. Thus, the method has great potential to be used as a screening method for the authentication of mammal and poultry species in meat products.

A rapid and straightforward classification of rice qualities or detection of food adulteration is necessary to meet the increasing demand of high quality rice, and to protect the consumers and supply chains from food fraud. Raman spectroscopy coupled with chemometrics have been used for multivariate analysis of rice quality and fraud detection. Supervised Kohonen Map (SKM) can classify different rice samples with low errors of Venetian-Blind (= 0.04) and Monte-Carlo (= 0.05) cross validation using the Raman spectral region of 200–1600 cm⁻¹. The classification performance of the FT-IR was examined and compared with those of Raman. For comparison, principal component analysis–linear discriminant analysis (PCA-LDA), classification and regression trees (CART), soft independent modeling by class analogy (SIMCA), and partial least squares-discriminant analysis (PLS-DA) techniques were also used for both Raman and FT-IR spectra. The top-5 classification models are “SKM + multiplicative scatter correction (MSC)” > “SKM + standard normal variate (SNV)” ~ “CART + MSC” > “SIMCA + MSC” > “SIMCA + SNV”. The proposed procedure showed better results than previous studies which can help both the industry and regulatory quality control to rapidly detect rice integrity and food fraud.

This study aimed to evaluate the effect of training-based intervention on process hygiene, food safety knowledge, and behavior of abattoir workers in Ethiopia. A total of 114 eligible participants and 138 swab samples from abattoir facilities were used to collect data. The pre- and post-intervention food safety knowledge and behaviors of participants were assessed using a structured questionnaire and direct observation, respectively. The swab samples were screened for hygiene indicator bacteria such as Escherichia coli, coliform, total coliform, Enterobacteriaceae, and aerobic plate count using petrifilm plates. The findings showed that participants’ food safety knowledge about pathogens and its associated illness (p = 0.004) and hygiene practices (p = 0.009) were significantly improved after intervention. The participant’s significant food safety behavioral change was observed in handwashing practices (p < 0.05). The participant’s behavior towards cleaning of work and meat contact surfaces was significantly improved after intervention (p = 0.000). Interestingly, the contamination level of generic E. coli (p = 0.034) and Enterobacteriaceae (p = 0.046) was significantly decreased in abattoirs after intervention. A significant reduction of generic E. coli contamination on beef carcasses (p = 0.009) and equipment (p = 0.036) was observed. The coliform (p = 0.013) and total coliform (p = 0.015) contamination of beef carcasses was also significantly reduced after intervention. Moreover, the personnel’s clothes and hands showed significantly lower Enterobacteriaceae contamination (p = 0.007) post intervention. The food safety training resulted in improvements of the hygiene process, some behaviors, and knowledge of participants. However, the implementation of integrated mitigation strategies is needed to ensure meat safety in abattoirs.

In many developing countries, the market for green foods is still nascent, and factors that influence their consumption are poorly understood. The present study aimed to investigate what drives the regular and occasional purchase behaviors of Iranian consumers. A systematic sampling technique was used to investigate the adoption behavior of 370 regular and occasional consumers in Karaj County, Iran. Also, a Bayesian model was constructed based on the theory of planned behavior, the value-belief-norm theory, and media influence. The Bayesian structural equation modeling revealed that the consumption behavior of both regular and occasional buyers appears to be driven by their environmental attitudes, environmental concerns, and also product accessibility. However, the relative effects of these determinants were different. In addition, the consumption behavior of regular buyers was influenced by green purchase perception, price fairness, purchase intention, and the media. Whereas other major determinants of occasional buyers’ behavior included hedonic attitude and labeling confidence. Recommendations and implications for policymakers and marketers were offered to increase green food consumption.