Daniella Ranario
Research Mentor(s): Liang Zhao
Mentor Department: CLaSP
Authors: Daniella Ranario, Liang Zhao, Yusuf Shwket
Session: Session 5 (2:00pm – 2:50pm)
Presentation Type: Poster 54
Abstract
Solar wind, a stream of charged particles released from the Sun’s atmosphere, plays a significant role in shaping space weather and impacts critical systems such as satellite operations, communication networks, and power grids. This research focuses on leveraging machine learning (ML) and artificial intelligence (AI) techniques to analyze heliophysics data and better understand the solar wind properties. Our initial efforts involved graphing and analyzing solar wind parameters—such as density, velocity, and composition ratios—to uncover underlying relationships within the data. Through this process, we identified notable outlier groups, including a distinct “tail” in the data corresponding to instances where the solar wind density falls below 1 cm-3. These observations highlight the importance of using advanced statistical and computational tools to explore such anomalies. Building on this foundation, we are integrating machine learning methodologies, including Principal Component Analysis (PCA), UMAP, clustering, and k-means, to systematically identify patterns, reduce dimensionality, and enhance data interpretation. This work will not only improve the predictive capabilities of solar wind models and contribute to more accurate space weather forecasting but also will establish a framework for future research to focus on specific properties or clusters, such as the observed tail, to gain deeper insights into solar wind behavior.
