-
Determination of class-specific variables in nonparametric multiple-class classification
Authors:
Wan-** Nicole Chen,
Yuan-chin Ivan Chang
Abstract:
As technology advanced, collecting data via automatic collection devices become popular, thus we commonly face data sets with lengthy variables, especially when these data sets are collected without specific research goals beforehand. It has been pointed out in the literature that the difficulty of high-dimensional classification problems is intrinsically caused by too many noise variables useless…
▽ More
As technology advanced, collecting data via automatic collection devices become popular, thus we commonly face data sets with lengthy variables, especially when these data sets are collected without specific research goals beforehand. It has been pointed out in the literature that the difficulty of high-dimensional classification problems is intrinsically caused by too many noise variables useless for reducing classification error, which offer less benefits for decision-making, and increase complexity, and confusion in model-interpretation. A good variable selection strategy is therefore a must for using such kinds of data well; especially when we expect to use their results for the succeeding applications/studies, where the model-interpretation ability is essential. hus, the conventional classification measures, such as accuracy, sensitivity, precision, cannot be the only performance tasks. In this paper, we propose a probability-based nonparametric multiple-class classification method, and integrate it with the ability of identifying high impact variables for individual class such that we can have more information about its classification rule and the character of each class as well. The proposed method can have its prediction power approximately equal to that of the Bayes rule, and still retains the ability of "model-interpretation." We report the asymptotic properties of the proposed method, and use both synthesized and real data sets to illustrate its properties under different classification situations. We also separately discuss the variable identification, and training sample size determination, and summarize those procedures as algorithms such that users can easily implement them with different computing languages.
△ Less
Submitted 7 May, 2022;
originally announced May 2022.
-
Single-crystal epitaxial europium iron garnet films with strain-induced perpendicular magnetic anisotropy: structural, strain, magnetic, and spin transport properties
Authors:
M. X. Guo,
C. K. Cheng,
Y. C. Liu,
C. N. Wu,
W. N. Chen,
T. Y Chen,
C. T. Wu,
C. H. Hsu,
S. Q. Zhou,
C. F. Chang,
L. H. Tjeng,
S. F. Lee,
C. F. Pai,
M. Hong,
J. Kwo
Abstract:
Single-crystal europium iron garnet (EuIG) thin films epitaxially strain-grown on gadolinium gallium garnet (GGG)(100) substrates using off-axis sputtering have strain-induced perpendicular magnetic anisotropy (PMA). By varying the sputtering conditions, we have tuned the europium/iron (Eu/Fe) composition ratios in the films to tailor the film strains. The films exhibited an extremely smooth, part…
▽ More
Single-crystal europium iron garnet (EuIG) thin films epitaxially strain-grown on gadolinium gallium garnet (GGG)(100) substrates using off-axis sputtering have strain-induced perpendicular magnetic anisotropy (PMA). By varying the sputtering conditions, we have tuned the europium/iron (Eu/Fe) composition ratios in the films to tailor the film strains. The films exhibited an extremely smooth, particle-free surface with roughness as low as 0.1 nm as observed using atomic force microscopy. High-resolution x-ray diffraction analysis and reciprocal space maps showed in-plane epitaxial film growth, very smooth film/substrate interface, excellent film crystallinity with a small full width at half maximum of 0.012$^{\circ}$ in the rocking curve scans, and an in-plane compressive strain without relaxation. In addition, spherical aberration-corrected scanning transmission electron microscopy showed an atomically abrupt interface between the EuIG film and GGG. The measured squarish out-of-plane magnetization-field hysteresis loops by vibrating sample magnetometry in conjunction with the measurements from angle-dependent x-ray magnetic dichroism demonstrated the PMA in the films. We have tailored the magnetic properties of the EuIG thin films, including saturation magnetization ranging from 71.91 to 124.51 emu/c.c. (increase with the (Eu/Fe) ratios), coercive field from 27 to 157.64 Oe, and the strength of PMA field ($H_\bot$) increasing from 4.21 to 18.87 kOe with the in-plane compressive strain from -0.774 to -1.044%. We have also investigated spin transport in Pt/EuIG bi-layer structure and evaluated the real part of spin mixing conductance to be $3.48\times10^{14} Ω^{-1}m^{-2}$. We demonstrated the current-induced magnetization switching with a low critical switching current density of $3.5\times10^6 A/cm^2$, showing excellent potential for low-dissipation spintronic devices.
△ Less
Submitted 11 January, 2022;
originally announced January 2022.
-
Fast Multi-Class Probabilistic Classifier by Sparse Non-parametric Density Estimation
Authors:
Wan-** Nicole Chen,
Yuan-chin Ivan Chang
Abstract:
The model interpretation is essential in many application scenarios and to build a classification model with a ease of model interpretation may provide useful information for further studies and improvement. It is common to encounter with a lengthy set of variables in modern data analysis, especially when data are collected in some automatic ways. This kinds of datasets may not collected with a sp…
▽ More
The model interpretation is essential in many application scenarios and to build a classification model with a ease of model interpretation may provide useful information for further studies and improvement. It is common to encounter with a lengthy set of variables in modern data analysis, especially when data are collected in some automatic ways. This kinds of datasets may not collected with a specific analysis target and usually contains redundant features, which have no contribution to a the current analysis task of interest. Variable selection is a common way to increase the ability of model interpretation and is popularly used with some parametric classification models. There is a lack of studies about variable selection in nonparametric classification models such as the density estimation-based methods and this is especially the case for multiple-class classification situations. In this study we study multiple-class classification problems using the thought of sparse non-parametric density estimation and propose a method for identifying high impacts variables for each class. We present the asymptotic properties and the computation procedure for the proposed method together with some suggested sample size. We also repost the numerical results using both synthesized and some real data sets.
△ Less
Submitted 4 January, 2019;
originally announced January 2019.
