Advancing Honey Authentication and Traceability Using Fluorescence Excitation-Emission Spectroscopy Coupled with Machine Learning Molecular Fingerprint Analysis

我们的创新项目旨在打击普遍存在的蜂蜜掺假问题，蜂蜜掺假已成为仅次于牛奶的全球第二大食品欺诈形式。这一令人担忧的问题破坏了消费者的信任，损害了蜂蜜行业的诚信，并带来了潜在的健康风险。欧盟报告《来自蜂巢》（From the hive）最近发表的研究结果显示，进口到欧盟的所有蜂蜜中，近一半含有外来糖源的标记，表明掺假。此外，从英国进口的蜂蜜的怀疑率为100%，这使英国蜂蜜的声誉受到威胁。我们将引入一种快速可靠的方法，将荧光激发发射（FLE）光谱与机器学习相结合，从而彻底改变蜂蜜的质量控制。目前的技术，如色谱法、核磁共振和感官分析，都是昂贵的、耗时的，而且正在过时。通过将现有技术FLE光谱应用于蜂蜜质量控制的新领域，并与机器学习相结合，我们对传统方法采取了颠覆性的方法。该项目利用先进的机器学习技术，如平行因子分析（PARAFAC）和偏最小二乘判别分析（PLS-DA）来分析蜂蜜样本。这些技术使我们能够确定样品中的化学成分、比例和质量标记。通过利用FLE光谱与机器学习算法相结合的速度，准确性和光学能力，我们的项目旨在超越目前最先进的蜂蜜质量评估方法。结果是更快、更可靠、可重复和成本效益高的质量控制方法，允许第三方验证产品内容。我们的联盟由经验丰富的养蜂人和专门从事蜂蜜生产、发光检测和数据分析的大学科学家组成，我们将利用他们超过25年的集体专业知识。通过这次合作，我们将开发创新、高效、经济的检测方法，准确检测蜂蜜掺假。实施我们提出的蜂蜜质量控制方法不仅将使我们的产品在国内和国际测试和分析服务市场上脱颖而出，而且还将建立新的产品质量和真实性测试标准套件。这将增强消费者信心，增加产品价值，并最终有助于保护英国蜂蜜的声誉。

Our innovative project aims to combat the widespread issue of honey adulteration, which has become the second most prevalent form of food fraud globally, following only milk. This concerning problem undermines consumer trust, compromises the integrity of the honey industry, and poses potential health risks. Recent findings published in the EU report "From the Hives" revealed that nearly half of all honey imported into the EU contained markers of extraneous sugar sources, indicating adulteration. Furthermore, a 100% suspicion rate was found for honey imported from the UK, putting the reputation of British honey at risk.We will revolutionize honey quality control by introducing a fast and reliable method using FLuorescence Excitation-Emission (FLE) spectroscopy combined with machine learning. Current techniques, such as chromatography, NMR and sensory analysis, are expensive, time-consuming and becoming outdated. By applying FLE spectroscopy, an existing technology, in a new area of honey quality control integrated with machine learning, we take a disruptive approach to conventional methods.The project utilizes advanced machine learning techniques like Parallel Factor Analysis (PARAFAC) and Partial Least Squares Discriminant Analysis (PLS-DA) to analyse honey samples. These techniques enable us to determine the chemical components, ratios, and quality markers in the samples. By harnessing the speed, accuracy, and optical capabilities of FLE spectroscopy in combination with machine learning algorithms, our project aims to surpass current state-of-the-art methods for honey quality assessment. The result is faster, more reliable, reproducible, and cost-effective quality control methods that allow for third-party verification of product contents.Our consortium is a collaboration of experienced beekeepers and university scientists specialising in honey production, luminescent detection, and data analytics, we will leverage their collective expertise of over 25 years. Through this collaboration, we will develop innovative, efficient, and cost-effective testing methods that accurately detect honey adulteration.Implementing our proposed approach to honey quality control will not only set our product apart from competitors in the national and international testing and analysis services markets but also lead to the establishment of new standard kits for product quality and authenticity testing. This will enhance consumer confidence, increase product value, and ultimately contribute to protecting the reputation of British honey.