Ensemble Based Temporal Weighting and Pareto Ranking (ETP) Model for Effective Root Cause Analysis

Root-cause identification plays a vital role in business decision making by providing effective future directions for the organizations. Aspect extraction and sentiment extraction plays a vital role in identifying the root-causes. This paper proposes the Ensemble based temporal weighting and pareto ranking (ETP) model for Root-cause identification. Aspect extraction is performed based on rules and is followed by opinion identification using the proposed boosted ensemble model. The obtained aspects are validated and ranked using the proposed aspect weighing scheme. Pareto-rule based aspect selection is performed as the final selection mechanism and the results are presented for business decision making. Experiments were performed with the standard five product benchmark dataset. Performances on all five product reviews indicate the effective performance of the proposed model. Comparisons are performed using three standard state-of-the-art models and effectiveness is measured in terms of F-Measure and Detection rates. The results indicate improved performances exhibited by the proposed model with an increase in F-Measure levels at 1%–15% and detection rates at 4%–24% compared to the state-of-the-art models.     

Al₂O₃ and γAl₂O₃ Nanomaterials Based Nanofluid Models with Surface Diffusion: Applications for Thermal Performance in Multiple Engineering Systems and Industries

Thermal transport investigation in colloidal suspensions is taking a significant research direction. The applications of these fluids are found in various industries, engineering, aerodynamics, mechanical engineering and medical sciences etc. A huge amount of thermal transport is essential in the operation of various industrial production processes. It is a fact that conventional liquids have lower thermal transport characteristics as compared to colloidal suspensions. The colloidal suspensions have high thermal performance due to the thermophysical attributes of the nanoparticles and the host liquid. Therefore, researchers focused on the analysis of the heat transport in nanofluids under diverse circumstances. As such, the colloidal analysis of H₂O composed by γAl₂O₃ and Al₂O₃ is conducted over an elastic cylinder. The governing flow models of γAl₂O₃/H₂O and Al₂O₃/H₂O is reduced in the dimensionless form by adopting the described similarity transforms. The colloidal models are handled by implementing the suitable numerical technique and provided the results for the velocity, temperature and local thermal performance rate against the multiple flow parameters. From the presented results, it is shown that the velocity of Al₂O₃–H₂O increases promptly against a high Reynolds number and it decreases for high-volume fraction. The significant contribution of the volumetric fraction is examined for thermal enhancement of nanofluids. The temperature of Al₂O₃–H₂O and γAl₂O₃–H₂O significantly increases against a higher ϕ. Most importantly, the analysis shows that γAl₂O₃–H₂O has a high local thermal performance rate compared to Al₂O₃–H₂O. Therefore, it is concluded that γAl₂O₃–H₂O is a better heat transfer fluid and is suitable for industrial and technological uses.     

Improved Epitaxy of AlN Film for Deep‐Ultraviolet Light‐Emitting Diodes Enabled by Graphene

The growth of single‐crystal III‐nitride films with a low stress and dislocation density is crucial for the semiconductor industry. In particular, AlN‐derived deep‐ultraviolet light‐emitting diodes (DUV‐LEDs) have important applications in microelectronic technologies and environmental sciences but are still limited by large lattice and thermal mismatches between the epilayer and substrate. Here, the quasi‐van der Waals epitaxial (QvdWE) growth of high‐quality AlN films on graphene/sapphire substrates is reported and their application in high‐performance DUV‐LEDs is demonstrated. Guided by density functional theory calculations, it is found that pyrrolic nitrogen in graphene introduced by a plasma treatment greatly facilitates the AlN nucleation and enables fast growth of a mirror‐smooth single‐crystal film in a very short time of ≈0.5 h (≈50% decrease compared with the conventional process), thus leading to a largely reduced cost. Additionally, graphene effectively releases the biaxial stress (0.11 GPa) and reduces the dislocation density in the epilayer. The as‐fabricated DUV‐LED shows a low turn‐on voltage, good reliability, and high output power. This study may provide a revolutionary technology for the epitaxial growth of AlN films and provide opportunities for scalable applications of graphene films.     

消费者属性对产品造型相似度认知的影响

目的 研究消费者属性对产品造型相似度认知的影响,为优化产品造型空间,提高人们对其认知的一致性提供依据。方法 首先,采用传统的方法,筛选出四十三个家用加湿器样本,构建产品造型空间,邀请七十三位被试,根据造型相似度将其进行多层分类,获取样本间造型相似度数据及被试属性数据;其次,基于轮廓分析法,依次分析不同属性下各组被试对各样本造型相似度认知的一致性;最后,根据异常样本的分布情况,分析各属性对产品造型相似度认知的影响程度。结论 生活形态属性对被试产品造型相似度认知的影响显著,异常率达20%以上,并且强于人口综合特征的影响;学历、性别、专业背景则基本无影响。应引入生活形态因素,检验人们对产品样本造型认知的一致性,进一步优化产品造型空间。     

Theory and practice of cushion curve: A supplementary discussion

The practice of developing cushion curves has been applied for many years. Known for its tedious development, the process does not correlate to cushion configurations too. This study reviewed conventional theories, practice of cushion curves, and simplified methods, including the conversion from quasi‐static compression chart to cushion curves, and testing‐extrapolated methods based on several applicable tests. In addition, the study examined the effect of configuration on cushion curves. Discussion indicated that the theories of cushion curves are based on two fundamental properties in physics: energy conversion and Newton's second law. Energy density‐based test‐extrapolation approaches reduced the development time significantly without compromising the accuracy of the cushion curves. The edge and corner cushion structure showed a similar pattern to the conventional flat cushion curve, but were shifted horizontally to the right. This paper recommends packaging engineers to apply C‐e curves as the basis of the cushion curves to design cushioning. Finally, energy dissipated during platen drop test is discussed to explain the deformation process in relation to dynamic stress‐strain curve and cushion curve.     

Co-Evolution in the Successful Learning of Backgammon Strategy

Following Tesauro's work on TD-Gammon, we used a 4,000 parameter feedforward neural network to develop a competitive backgammon evaluation function. Play proceeds by a roll of the dice, application of the network to all legal moves, and selection of the position with the highest evaluation. However, no backpropagation, reinforcement or temporal difference learning methods were employed. Instead we apply simple hillclimbing in a relative fitness environment. We start with an initial champion of all zero weights and proceed simply by playing the current champion network against a slightly mutated challenger and changing weights if the challenger wins. Surprisingly, this worked rather well. We investigate how the peculiar dynamics of this domain enabled a previously discarded weak method to succeed, by preventing suboptimal equilibria in a meta-game of self-learning.     

基于空间变换密集卷积网络的图片敏感文字识别

互联网上含有大量多字体混合、形变、拉伸、左右结构字形、倾斜畸变等复杂场景下的敏感文字图片,在处理相关图片过程中存在特征提取难、识别率低的问题.本文提出基于空间变换网络与密集神经网络的方法对图片敏感文字进行特征提取与变换矫正,使用了深层双向GRU网络与CTC时域连接网络对序列特征信息进行标记预测,序列化处理文本的方式可较好地提升距离较宽文字与模糊文字信息的处理能力.实验结果表明,本模型在Caffe-OCR中文合成数据集和CTW数据集中分别实现了87.0%和90.3%识别准确率,平均识别时间达到了26.3 ms/图.     

视频监控场景下基于单视角步态的人体身份及属性识别系统

基于步态的特征识别是一种新兴的生物特征识别技术, 旨在通过人们走路的姿态进行身份和相关属性的分析. 与其他的生物识别技术相比, 基于步态的识别方法具有难隐藏性、非接触性和可远距离使用的优点. 本文设计出一个视频监控场景下基于单视角步态的人体身份及属性识别系统, 该系统通过图像处理方法从复杂的监控视频中实时检测出人体的步态, 经过利用深度学习训练过的算法进行分析后, 获取人体的身份、性别和年龄信息. 实验表明, 系统的身份识别准确率达98.1%, 性别预测准确率达97.1%, 年龄预测平均绝对误差为6.21岁, 实验结果均优于传统基准算法, 且系统开发成本低, 支持实时检测, 能充分满足中小规模步态研究与分析的需要.