Title :
Scalable Multivariate Time-Series Models for Climate Informatics
Author_Institution :
Univ. of Southern California, Los Angeles, CA, USA
Abstract :
The increasing volume of climate data has created the need for scientists to develop scalable data analysis tools beyond traditional techniques. Climate data not only have a massive scale but also high dimension and complex dependency structures, making the analysis task extremely challenging. Climate informatics leverages advanced algorithmic tools from data science to solve problems in climate science. This article showcases how scalable multivariate time-series models can be developed for climate change attribution, spatiotemporal analysis, and extreme value time-series analysis.
Keywords :
climatology; data analysis; geophysics computing; time series; climate change attribution; climate data; climate informatics; data analysis; extreme value time-series analysis; multivariate time-series models; spatiotemporal analysis; Analytical models; Computational modeling; Data models; Hidden Markov models; Meteorology; Spatiotemporal phenomena; Tensile stress; data mining; knowledge management applications; machine learning; scientific computing; time series analysis;
Journal_Title :
Computing in Science Engineering
DOI :
10.1109/MCSE.2015.126
