A Novel Approach for Exploring Data-Driven Nutritional Insights Using Clustering and Dimensionality Reduction Techniques

The analysis of high-dimensional datasets poses significant challenges, particularly in big data analytics where extracting meaningful insights is crucial. Current techniques often struggle with maintaining a balance between preserving global data structures and capturing local relationships. In this study, we address these challenges by integrating Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE) for dimensionality reduction, with a dedicated focus on the domain of nutrition. Our research specifically targets pressing health concerns among the youth, such as obesity and nutritional deficiencies, by simplifying the analysis of extensive nutrition datasets. We propose a methodology that enhances the selection of nutritious food alternatives through effective data simplification and analysis. The efficacy of our approach is demonstrated through comprehensive experimental results, which include detailed comparisons with state-of-the-art methods, and evaluations based on clustering accuracy, computational efficiency, and visualization quality. Additionally, we optimize the performance of clustering algorithms using hyperparameter tuning techniques, specifically the Elbow Method and the Silhouette Coefficient. Our findings highlight the significant role of dimensionality reduction in improving data analysis and machine learning processes. This study offers valuable insights for researchers and practitioners, contributing to a deeper understanding of how dimensionality reduction techniques can unlock latent knowledge within vast datasets. Ultimately, our research aims to facilitate more informed decision-making and drive innovation in the era of big data analytics, with practical applications extending across diverse domains, particularly in nutrition and health.