We present comprehensive studies of strain effects on the spin reorientation transition (SRT), the so-called Morin transition, in α-Fe2O3(0001) films. The α-Fe2O3(0001) epitaxial films were grown with a Cr2O3 buffer layer on Al2O3(0001) substrates through an oxide molecular beam epitaxy. The antiferromagnetic spin axis was monitored by using the Fe L2-edge X-ray magnetic linear dichroism. The buffer layer was found to introduce compressive strain into the α-Fe2O3(0001) film due to its 1.6% smaller in-plane lattice constant. The degree of strain is monotonically reduced with the increase of the α-Fe2O3 film thickness and becomes relaxed in the thick region (>20 nm). The transition temperature TM , which increases up to 360 K, well above the bulk TM = 263 K, for the film thickness 3 nm, gradually decreases as the film thickness increases. We also examined the Néel temperature TN in the ultra-thin region (<3 nm), which rapidly drops with the decrease of the film thickness. The correlation between TM and the strain in the α-Fe2O3(0001) epitaxial films was found to be well explained in terms of two competing energies of magnetic dipole anisotropy and single-ion magnetocrystalline anisotropy except for the ultra-thin region, in which TN is dominated by the finite-size effects.

Summary As ontologies are becoming larger and more diverse, ontological analysis and visualization of results have become more challenging, rendering the need for more computing resources. To address this issue, we suggest a system based on Hadoop for ontological analysis for large ontologies. Our suggested system consists of three parts: a data server to analyze ontological data, a visualization server to visualize the result of data analysis, and user applications to provide users with the visualized data. Server applications are implemented based on the Hadoop framework, and the ontological data are processed efficiently using the MapReduce algorithm. We performed ontological analysis using the DBpedia ontology and visualized the result. The goal of the visualization process is to determine the major properties of each class, and visualization is conducted on the Web in order to provide users with a cross‐platform environment. We evaluate the performance of the method by measuring execution times and analyzing experimental results obtained in the visualization process. The system we present is scalable for big ontological data. Copyright © 2015 John Wiley & Sons, Ltd.

No abstract

No abstract

Detecting differences between XML documents is one of most important research topics for XML. Since XML documents are generally considered to be organized in a tree structure, most previous research has attempted to detect differences using tree-matching algorithms. However, most tree-matching algorithms have inadequate performance owing to limitations in terms of the execution time, optimality and scalability. This study proposes a stream-based difference detection method in which an XML binary encoding algorithm is used to provide improved performance relative to that of previous tree-matching algorithms. A tree-structured analysis of XML is not essential in order to detect differences. We use a D-Path algorithm that has an optimal result quality for difference detection between two streams and has a lower time complexity than tree-based methods. We then modify the existing XML binary encoding method to tokenize the stream and the algorithm in order to support more operations than D-Path algorithm does. The experimental results reveal greater efficiency for the proposed method relative to tree-based methods. The execution time is at least 4 times faster than state-of-the-art tree-based methods. In addition, the scalability is much more efficient.