[M-GI27] データ駆動地球惑星科学

[MGI27-P09] 固体地球科学分野におけるデータ駆動型解析の進展と今後の展望

*桑谷 立^1,2、岡本 敦³、吉田 健太¹、中村 謙吾³、土屋 範芳³、駒井 武³ (1.国立研究開発法人 海洋研究開発機構、2.国立研究開発法人 科学技術振興機構、3.東北大学大学院環境科学研究科)

キーワード：データ駆動型解析、固体地球、ベイズ推定、疎性モデリング

The high-dimensional and large amounts of data sets in geosciences show very complex behavior and often have large uncertainty. It is important to extract a small number of essential parameters which can explain the phenomenon from high-dimensional data in order to understand the behavior of dynamic solid earth. Under the framework of a big scientific project entitled as “Initiative for high-dimensional data-driven science through deepening sparse modelling” supported by the MEXT in Japan ( http://sparse-modeling.jp/index_e.html ), we try to introduce data-driven approaches into geosciences. In this presentation, we will share some applications in solid-earth science and discuss future prospects.