The demand for meat products has increased significantly due to changes in people’s lifestyles and rising meat prices. The study aims to investigate the potential use of an accurate and reliable method for detecting adulteration and ensuring the quality of meat products. The sausages are made with various meat from 5 to 100% pure meat. The samples are then analyzed using Fourier-transform Near-Infrared (FT-NIR) spectroscopy. This research used a NIR spectrophotometer, Principal component analysis (PCA), and Linear discriminant analysis (LDA) to identify beef, chicken, donkey, and horse handmade sausage samples. Using PCA, two principal components were extracted that covered 99 percent of the data variance and accurately distinguished sausage samples from different animal meats. Following this, the LDA model can confidently classify the sausage samples and achieved 100% accuracy in classifying all test samples. To conclude, NIR spectroscopy is a reliable tool for sausage quality evaluation and species identification, which can identify different animal species in meat products.

Ensuring accurate meat species identification and animal authentication in meat products is crucial for promoting fair trade and empowering consumers to make informed decisions. Concerns about food fraud have grown due to significant financial and trust-related consequences. Chromatographic techniques combined with chemometrics are widely used in the food industry to confirm the origin and authenticity of food products. Liquid chromatography-mass spectrometry (LC-MS) is highly sensitive and can accurately identify meat metabolites and lipids. This systematic review focuses on analytical methods for the quantification and authentication of meat products. Data were collected from authoritative databases such as Web of Science, Scopus, and PubMed between January 2014 and February 2023. The review examines the use of LC-MS to identify the nature and origin of food. Studies demonstrate the high efficiency of LC-MS due to advanced techniques for extraction, concentration, and identification, enabling the separation and identification of trace amounts with high accuracy and specificity. Identifying peptide markers is crucial for developing these methods. LC-MS/MS stands out for its accuracy in detecting food adulteration, relying on the reliability of proteotypic peptides unique to the animal species used as food. The large amount of data generated during mass spectrometry necessitates the use of chemometric methods to analyze LC-Q-TOF data. The combination of LC-Q-TOF and chemometrics has become a powerful tool for the qualitative and quantitative analysis of food ingredients. The results indicate that LC-MS/MS proteomics tools effectively identify and specify species markers in various food types, including meat, delivering accurate and statistically robust results.

In recent years, consumers have become increasingly concerned about the safety and authenticity of their food, as it directly affects their health. Food adulteration is a global issue that poses safety risks and is difficult to detect. Economically motivated adulteration (EMA) frequently results from the diversion of ingredients and unauthorized consumption, making food authenticity an important issue requiring appropriate identification techniques. Advanced detection methods like high-performance liquid chromatography (HPLC) and gas chromatography-mass spectrometry (GC-MS) can identify adulterants but are limited by high costs and complex sample preparation. Therefore, fast, non-destructive, and accurate food adulteration detection techniques are essential. This review focuses on spectroscopic methods, including near-infrared (NIR), mid-infrared (Mid-IR), and Fourier transform infrared (FTIR) spectroscopy, for identifying food authenticity. Studies have shown that these methods, which offer rapid, non-invasive, and cost-effective physicochemical fingerprinting, have largely replaced classical analytical techniques, enhancing productivity and profitability in the food supply chain.

Relation extraction (RE) is a critical task in natural language processing that identifies relationships between entities within text. While increasing the size of annotated datasets generally enhances model performance, the cost and effort associated with human annotation can be prohibitive. This study investigates the impact of dataset size on RE performance using the Semantic Evaluation 2010 Task 8 dataset. Four deep neural models-Attention-Based Convolutional Neural Network, Attention-Based Bidirectional Long Short-Term Memory Network, Relation Bidirectional Encoder Representations from Transformers, and Matching the Blanks-were evaluated across eight subsets of the dataset to simulate size reduction. Performance metrics such as accuracy, precision, recall, and F1-score were analyzed. The results indicate that although larger datasets improve model performance, the benefits diminish beyond a certain size, illustrating the law of diminishing returns. These findings emphasize the importance of balancing dataset size and quality for cost-effective training and suggest strategies such as data augmentation and transfer learning to optimize performance with limited data.