An Optimization-Based Supervised Learning Algorithm for PXRD Phase Fraction Estimation

In powder diffraction data analysis, phase identification is the process of determining crystalline phases in a sample using its characteristic Bragg peaks. For multiphasic spectra, we must also determine relative weight fraction each sample. Machine learning algorithms (e.g., Artificial Neural Networks) have been applied to perform such difficult tasks but typically require significant number training samples for acceptable performance. We developed an approach that performs well even with small samples. apply fixed-point iteration algorithm on labelled estimate monophasic spectra. Then, given unknown spectrum, again use weighted combination monophase spectra best approximates spectrum. These weights are desired fractions Our gave accuracy and lowest mean absolute error when compared several machine algorithms, including single hidden-layer neural network.