Effect of annealing atmosphere on the structure and dielectric properties of unfilled tungsten bronze ceramics Ba4PrFe0.5Nb9.5O30

Unfilled tungsten bronze ceramics with the nominal formula Ba₄PrFe_0.5Nb_9.5O₃₀ were synthesized via the standard solid-state sintering route, and the effects of oxygen vacancies on the dielectric and electrical properties were investigated in addition to the structure. Room-temperature X-ray diffraction showed that the N₂-annealed sample had the largest cell volume. Low-temperature spectrum showed that N₂ annealing rendered the dielectric constant and dielectric loss more frequency dispersive, whereas O₂ annealing inhibited the frequency dispersion. The dc conductivity of all the samples originated from the electrons produced in the second ionization of oxygen vacancies and was most likely controlled by a mixed conduction mechanism of the electron and oxygen-vacancy ions. The N₂-annealed sample has the highest dc conductivity owing to its high concentration of oxygen vacancies. The broadening of the Raman lines and the decrease of Raman intensity for the N₂-annealed sample originated from a significant structural disorder. X-ray photoelectron spectra demonstrated that the increased oxygen vacancies caused by the change of valences of Fe and Pr ions contributed to the structural disorder.     

改进的多层感知机在客户流失预测中的应用

针对在传统的客户流失预测数据预处理中，使用one-hot编码处理离散属性导致数据维度增加及数据过于稀疏的问题，提出了两种基于多层感知机的改进后的客户流失预测模型。其主要思想是分别使用堆叠自编码器和实体嵌入两种方法对多层感知机进行改进，通过将离散属性的高维编码数据向低维空间映射，有效地减少了one-hot编码产生的稀疏数据，增加了离散属性值之间的关联度。在对两份公开的数据集进行交叉验证后的实验结果表明，改进后的模型既有效地提高了预测的准确度，又维持了传统多层感知机模型在并行化计算方面的优势。     

Global Exponential Stability of Hybrid Non-autonomous Neural Networks with Markovian Switching

Neural Processing Letters - This paper discusses the global exponential stability for a class of hybrid non-autonomous neural networks (HNNNs) with Markovian switching, which includes the factors...     

An image encryption scheme based on precision limited chaotic system

Multimedia Tools and Applications - In recent years, various chaotic maps have been used for image encryption. However, most of these image encryption algorithms entail a lot of floating-point...     

ICFR: An effective incremental collaborative filtering based recommendation architecture for personalized websites

World Wide Web - To solve the problem that users’ retrieval intentions are seldom considered by personalized websites, we propose an improved incremental collaborative filtering (CF)-based...     

A Pseudo-document-based Topical N-grams model for short texts

World Wide Web - In recent years, short text topic modeling has drawn considerable attentions from interdisciplinary researchers. Various customized topic models have been proposed to tackle the...     

基于智能电表大数据的异常用电检测

异常或欺诈造成的非技术性电力损失是电力公司损失的主要源头之一。智能电表的广泛使用，使得运用大数据方法实现对非技术性电力损失的有效检测成为可能。提出了一种使用监督学习进行非技术损失检测的方法。该方法基于智能仪表记录的所有信息（耗电量、异常警报等）结合辅助数据库所提供的有关每个智能电表的地理位置和技术参数的附加信息，使用最优的机器算法来深入分析用电客户的用电行为，生成异常用电客户列表。通过现场检查的结果表明，该方法能够较为准确地识别智能电网中所存在异常用电客户。     

PWR堆芯燃料组件非线性梁模型研究

为描述燃料组件的非线性特性，采用有限元方法建立燃料组件横向梁模型，并引入迟滞模型，用于模拟燃料组件横向变形时的非线性效应。计算结果表明:该非线性梁模型能够更准确地获得组件的弯曲变形和受力，有助于开展燃料组件事故分析。      

Low-complexity transmit antenna selection algorithm formassive MIMO

Massive multiple input multiple output (MIMO) systems can increase capacity and reliability greatly. However,extremely high hardware costs and computational complexity lead to the demand for reasonable antenna selection.Aiming at the problem that the traditional antenna selection algorithm based on maximizing sum capacity has largecomplexity and worse bit error rate (BER) performance, a two-step selection algorithm is proposed, which selectsa part of the antennas based on the norm-based antenna selection (NBS) firstly, and then selects the antenna basedon maximizing capacity via convex optimization. The simulation results show that the improved algorithm has betterBER performance than the traditional algorithms. At the same time, it reduces computational complexity greatly.