Exploratory data analysis (EDA) machine learning approaches for ocean world analog mass spectrometry

Many upcoming and proposed missions to ocean worlds such as Europa, Enceladus, Titan aim evaluate their habitability the existence of potential life on these moons. These will suffer from communication challenges technology limitations. We review investigate applicability data science unsupervised machine learning (ML) techniques isotope ratio mass spectrometry (IRMS) volatile laboratory analogs Europa Enceladus seawaters a case study for development new strategies icy world missions. Our driving goal is determine whether spectra gases could contain information about composition seawater biosignatures. implement ML what inherent pipeline be designed quickly analyze future In this study, we focus exploratory analysis (EDA) step in analytics pipeline. This crucial that allows us understand depth before subsequent steps predictive/supervised learning. EDA identifies characterizes recurring patterns, significant correlation structure, helps which variables are redundant contribute variation lower dimensional space. addition, identify irregularities outliers might due poor quality. compared dimensionality reduction methods Uniform Manifold Approximation Projection (UMAP) Principal Component Analysis (PCA) transforming our high-dimensional space dimension, clustering algorithms identifying data-driven groups (“clusters”) analog IRMS mapping clusters experimental conditions CO 2 concentration. Such characterization efforts first toward longer-term autonomy where similar automated tools used onboard spacecraft prioritize transmissions bandwidth-limited outer Solar System