E. Schmitt | Searcharxiv

arXiv2015

Robust PCA with FastHCS

Principal component analysis (PCA) is widely used to analyze high-dimensional data, but it is very sensitive to outliers. Robust PCA methods seek fits that are unaffected by the outliers and can therefore be trusted to reveal them. FastHCS (High-dimensional Congruent Subsets) is a robust PCA algorithm suitable for high-dimensional applications, including cases where the number of variables exceeds the number of observations. After detailing the FastHCS algorithm, we carry out an extensive simulation study and three real data applications, the results of which show that FastHCS is systematically more robust to outliers than state-of-the-art methods.

stat.ME↗

arXiv2005

Heat capacity of $α$-GaN: Isotope Effects

Until recently, the heat capacity of GaN had only been measured for polycrystalline powder samples. Semiempirical as well as \textit{first-principles} calculations have appeared within the past few years. We present in this article measurements of the heat capacity of hexagonal single crystals of GaN in the 20-1400K temperature range. We find that our data deviate significantly from the literature values for polycrystalline materials. The dependence of the heat capacity on the isotopic mass has also been investigated recently for monatomic crystals such as diamond, silicon, and germanium. Multi-atomic crystals are expected to exhibit a different dependence of these heat capacities on the masses of each of the isotopes present. These effects have not been investigated in the past. We also present \textit{first-principles} calculations of the dependence of the heat capacities of GaN, as a canonical binary material, on each of the Ga and N masses. We show that they are indeed different, as expected from the fact that the Ga mass affects mainly the acoustic, that of N the optic phonons. It is hoped that these calculations will encourage experimental measurements of the dependence of the heat capacity on isotopic masses in binary and more complex semiconductors.

cond-mat.mtrl-sci↗