Training flowchart and inference pipeline for GC-MS dataset — astro-ph.EP
This paper presents the application of artificial intelligence to mass spectrometry data to detect the potential habitability of ancient Mars.
Although the data was collected for the planet Mars, the same approach can be replicated for any terrestrial object in our solar system. Moreover, the proposed methodology can be adapted to any domain where mass spectrometry is used. This research focuses on data analysis of two mass spectrometry techniques, namely evolved gas analysis (EGA-MS) and gas chromatography (GC-MS), which are used to identify certain chemical compounds in geological material samples.
This study shows the application of EGA-MS and GC-MS data for the analysis of extra-terrestrial materials. The most important features of the proposed methodology include square root transformation of mass spectrometry values, conversion of raw data into 2D spectrograms, and utilization of specific machine learning models and techniques to avoid overfitting on relatively small data sets.
The EGA-MS and GC-MS data sets come from NASA and two machine learning competitions in which the authors participated and utilized. Complete running code for the GC-MS dataset/competition is available on GitHub.1 Raw training mass spectrometry data includes [0, 1] labels of specific chemical compounds, chosen to provide valuable insight and contribute to our understanding of the potential past habitability of Mars.
Ioannis Nasios
Subjects: Astrophysics of the Earth and Planets (astro-ph.EP); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Data Analysis, Statistics and Probability (physics.data-an)
Cite as: arXiv:2310.11888 [astro-ph.EP] (atau arXiv:2310.11888v1 [astro-ph.EP] for this version)
Related DOIs:
https://doi.org/10.1016/j.icarus.2023.115824
Focus on learning more
Delivery history
From: Ioannis Nasios
[v1] Rabu, Oct 18, 2023 11:14:46 UTC (1,453 KB)
https://arxiv.org/abs/2310.11888
Astrobiologi,
2023-10-19 21:18:28
#Analyzing #Mass #Spectrometry #Data #Artificial #Intelligence #Aid #Understanding #Mars #Habitability #Provide #Insight #Future #Missions #SurabayaPostKota.net
