2016 年 37 巻 12 号 p. 610-615
Recent development of digitized and automated measurement systems has inevitably accelerated applications of data mining techniques based on statistical/information processing. In the present article we outlined the recent progress in our chemical imaging techniques based on the statistical analysis of ‘spectral image’ datasets, obtained by a suite of scanning transmission electron microscopy and associated spectroscopy. Finally, we discuss the future prospects of the field to extend the present bilinear model to multi-way analysis for more robust modeling of low signal-to-noise ratio data.