摘要
大数据是"未来的新石油",Nature和Science相继出版专刊来探讨大数据带来的挑战和机遇。大数据的特征是数据规模大,并经常呈异构多模态、复杂关联、动态涌现等特点,需要高效计算模型和方法。大数据-智能矿床研究刚刚起步,需对多维、异构、隐性大数据的高效存储、管理、集成、融合与深度挖掘,需人工智能方法——机器学习、深度学习、可视分析的应用。贝叶斯网络是成因建模的一个革命性工具,可以用来揭示矿床的成因机制及它们背后的规律。来自地质调查、监测数据获得的与"矿"有关的大数据,通过迭代计算,可以不断完善所建立的矿床模型,并且通过云计算技术,使得世界各地的矿床研究团队共同参与,引发矿床模型研究方式的变革。
Big data is the key resource in the future. Both Nature and Science have published special issues to discuss the challenge and opportunity brought by big data. Big data are characterized by large scale, as well as by heterogeneous multi-modal, complicated interaction, dynamic emergence and others, so that highly efficient com- puting methods and models are required. Study on big data-based intelligent mineral deposit has just started, pushing for efficient storage and management, integration, fusion and deep mining of multi-dimensional, hetero- geneous and recessive big data, as well as for the application of artificial intelligence methods such as machine learning, in-depth learning and visual analysis. The Bayesian network is a revolutionary tool for genetic modeling. It can be used to reveal the genetic mechanism of mineral deposits and the patterns behind them. Mineral deposit- related big data from geological survey and monitoring data could constantly improve the deposit model through continuing iterative calculation. Furthermore, the cloud computing technology makes it possible to concurrently study a deposit model by research teams from all over the world at the same time, and thus leads to a revolution of the research mode of deposit model.
出处
《矿物岩石地球化学通报》
CSCD
北大核心
2017年第2期327-331,344,共6页
Bulletin of Mineralogy, Petrology and Geochemistry
基金
国家重点研发计划重点专项课题(2016YFC0600506)
关键词
大数据
智能矿床模型
机器学习
深度学习
贝叶斯网络
big data
intelligent modelling of mineral deposit
machine learning
in-deepth learning
Bayesian network