Abstact
Phase pattern decomposition of inorganic materials' crystalline
structure is extremely important for the unearthing of new properties such as superconductivity.
Previously, this process had meticulously been done by hand, so computer algorithms
have been developed to try and uncover these phases. They, however, have yet to
combine eciency and accuracy together. The goal of this project is to do just that
by extending the Graph-based endmember extraction and labeling algorithm (GRENDEL). P
hase one will be to incorporate physical constraints and prior knowledge to
increase the accuracy of our phase composition results, and phase two will be to utilize
active learning to minimize the number of sample points needed to analyze a given
material to increase eciency.