Research Summary：Many kinds of spatio-temporal data can be obtained from industries. All types of data, such as images, texts, and sensitivity data, as well as multidimensional high-frequency observations are available there. To solve actual problem, we have developed statistical models and examined those statistical properties.

Plant traits (flowering and fruiting, adaptability to the environment, amount of seeds, etc.) are determined by the interaction between genes encoded in the genome (genetic factors) and environmental changes such as light and temperature (environmental factors). Variable selection by sparse modeling is effective for identifying genes and environmental factors that are explanatory variables of phenotypes, and is useful for extracting valuable information for promoting precision agriculture and crop breeding efficiently.