Hello Everyone!

Food insecurity is a prominent issue in our world, with many people worldwide suffering from malnutrition due to a lack of essential nutrients in their bodies each day, attributed to the high-cost-high-nutrition foods and a lack of awareness. Currently, some soup kitchens and non-profit organizations are focused on delivering food items with superior nutritional content; however, their limited availability and high price put a limitation on the quantity for mass distribution. This leads to the consumption of easily available food items that are high in sugar and unhealthy fats. The goal of this study is to locate the easily available low-cost options from all the essential food clusters. Two different unsupervised machine learning algorithms, i.e., K-means Clustering and hierarchical clustering, were applied to the frequently consumed food items in most of the houses. The output was three clusters of food items based on the degree of their nutritional significance. After analysis of these 3 clusters, the food item that is low-priced can be selected from each cluster, since they deliver the same benefit as their expensive counterparts but at an affordable price band. For example, a cluster of similar nutritional benefits contains expensive foods, such as high-end cheese, compared to rice dishes, which are only a fraction of the cost. In the future, this project hopes to expand by using a larger dataset and creating more clusters to find affordable swaps for organizations to optimize the cost-nutrition ratio.

Due to rapid growth rates of industrialization and urbanization, water resources are extensively contaminated, thus triggering the need to develop sustainable wastewater treatment materials at low costs. Biochar was produced from agro-waste orange peels using pyrolysis and tested as a prospective water treatment technology to separate heavy metals from wastewater. The surface morphological properties of the synthesized biochar were analyzed by FESEM, which clearly identified that the surface of the biochar had formed a heterogeneous structure post-pyrolysis, and the surface groups were identified by Fourier Transform Infrared Spectrometry (FTIR). Adsorption studies were performed to assess the efficiency of lead (Pb), cadmium (Cd), copper (Cu), and zinc (Zn) metal ion removal. The biochar sample obtained from 400°C showed considerable reduction in metal ion concentration, which removed 59.8\% of Cu, 56.6\% of Cd, 50.3\% of Pb, and 63.9\% of Zn metal ion from the solutions. Several factors like contact time, dosage, and pH of the solutions were found to affect the adsorption capacity.

From these results, it can thus be concluded that biochar derived from orange peels is an effective, eco-friendly, and economically viable adsorbent. Through this study, the importance of agro-waste management is recognized.

The move towards a circular economy demands integrated approaches that integrate nutritional security, agricultural waste conversion, and sustainable material innovation. In this study, a decision support system based on AI is presented to select food resources that offer nutritional value and produce high-quality fibrous by-products for composite material synthesis. A knowledge base was created to associate food resources with nutritional properties, waste generation routes, and economic viability. These qualitative variables were converted into quantitative scores for nutrition value, waste-to-fiber potential, and economic viability.

Machine learning algorithms were utilized to process and optimize these variables. XGBoost regression was utilized for feature interpretation and model explainability, and a neural network was utilized to understand non-linear relationships to produce optimized rankings for a circular economy. SHAP values were used for transparent and interpretable model explanation. Furthermore, Pareto-based multi-objective optimization was used to maximize nutritional value and waste valorization potential while minimizing economic costs, providing optimal trade-off solutions.

To validate the proposed approach, a coir-flax fiber reinforced epoxy composite was produced using the selected fibrous waste resources. Mechanical testing confirmed a tensile strength of 75 MPa and a tensile modulus of 7.5 GPa, establishing the structural viability of AI-assisted material selection.