science_cases:lmsu_science_case
Differences
This shows you the differences between two versions of the page.
Both sides previous revisionPrevious revisionNext revision | Previous revision | ||
science_cases:lmsu_science_case [2020/11/04 11:58] – admin | science_cases:lmsu_science_case [2022/09/12 13:44] (current) – admin | ||
---|---|---|---|
Line 3: | Line 3: | ||
===== Short description of the physical problem ===== | ===== Short description of the physical problem ===== | ||
- | <font 11ptfont-family: | + | The interplanetary medium is filled with the solar wind and the interplanetary magnetic field (IMF). |
- | <font 11ptfont-family: | + | The magnetopause |
- | <font 11ptfont-family: | + | IMF pressure + solar wind plasma pressure = planetary and magnetospheric |
- | <font 11ptfont-family: | + | In the case of Mercury, in front of the magnetopause the solar wind ram pressure, and inside it the magnetospheric magnetic field pressure play the main roles. Mercury' |
+ | ===== Aim of the science case ===== | ||
- | <font 11ptfont-family: | + | The MESSENGER (MErcury Surface, Space ENvironment, |
- | <font 11ptfont-family: | + | During each orbit, MESSENGER typically spent 1–2 h inside the magnetosphere; the rest of the time was spent in the magnetosheath and in the interplanetary medium. The interplanetary magnetic field (IMF) magnitude assigned to each crossing can be evaluated as a 1 hour average of magnetometer (MAG) data upstream |
- | ===== Aim of the science case ===== | + | |
+ | Magnetic field data are analyzed in Mercury | ||
+ | |||
+ | The goal of this case is to improve our understanding of Mercury' | ||
+ | |||
+ | ===== Description of the machine learning problem and our approach ===== | ||
+ | |||
+ | Based on data from the mission, several global models of the magnetosphere were proposed (e.g., Winslow et al., 2013; Philpott et al., 2020). However, they could only describe an average shape of the bow shock and magnetopause crossings and can be prone to missing the statistical nuances in the data. Given large data, [[:glossary#neural_network|neural networks]] can be expected to approximate complex functions, which often surpass deterministic and rule-based methods, in a variety of time series tasks like [[:glossary#classification|classification]] (Fawaz et al., 2019), time series forecasting (Lim and Bohren, 2021), and rare time series event detection (Nguyen et al., 2018). We leverage these to **develop a predictor** that can be used in real-time during orbit to predict magnetic region for each step in a short window of observation. | ||
+ | |||
+ | The use of **statistical neural networks** allows us to explore another aspect: With the help of **[[:glossary#active_learning|active learning]]** | ||
+ | |||
+ | The data set was **manually labelled with the boundary crossings**. To identify | ||
+ | |||
+ | The distribution of the different regions, after annotation, is reported in the following table. | ||
+ | |||
+ | ^Label^Region^Statistical distribution (%)| | ||
+ | |0|Interplanetary magnetic field|65.4| | ||
+ | |1|Bow shock crossing|3.7| | ||
+ | |2|Magnetosheath|14.5| | ||
+ | |3|Magnetopause crossing|2.3| | ||
+ | |4|Magnetosphere|14.1| | ||
+ | |||
+ | The boundaries of critical interest - bow shock and magnetopause - are minorities with only 3.7 and 2.3 % representation. The table highlights | ||
+ | |||
+ | As a first step in pre-processing, | ||
+ | |||
+ | In the feature preparation stage, a sliding window of variable sizes (3 seconds to 3 minutes) | ||
+ | |||
+ | The windowed features are fed first into a block of 3 [[: | ||
- | <font 11ptfont-family: | + | The window size used in these experiments is 30 seconds. Overall, the [[:glossary#predictor|predictor]]** achieves a macro [[:glossary#f1|F1 score]] of about 80% **on the bow shock and the magnetopause crossings on a randomly sampled test of 300 orbits. None of the orbits overlap in the train and test sets. |
- | <font 11ptfont-family:Arial;color:# | + | Results of this science case were presented at the {{:wiki:egu2021-lavrukhin_etal.pdf|EGU21}} |
- | <font 11ptfont-family:Arial; | + | **References:** |
- | <font 11ptfont-family:Arial;color:# | + | * Philpott, L.C., et al. (2020), The Shape of Mercury’s Magnetopause: The Picture From MESSENGER Magnetometer Observations and Future Prospects for BepiColombo, |
+ | * Winslow, R. M., et al. (2013), Mercurys magnetopause and bow shock from MESSENGER Magnetometer observations, | ||
+ | * Fawaz, H.I., et al. (2019), Deep learning for time series classification: a review, Data Mining and Knowledge Discovery 33, doi: 10.1007/s10618-019-00619-1 | ||
+ | * Lim, B., and Zohren, S. (2021), Time-series forecasting with deep learn- ing: a survey, Philosophical Transactions of the Royal Society A: Mathematical, | ||
+ | * Nguyen, V., et al. (2018), Applications | ||
science_cases/lmsu_science_case.1604487538.txt.gz · Last modified: 2020/11/04 11:58 by admin