Sample size and error in the determination of mode shapes by principal components analysis

Principal components analysis (PCA) is a multivariate statistical technique that transforms a data set having a large number of inter-related variables to a new set of uncorrelated variables called the principal components, determined to allow the dimensionality of the data set to be reduced while retaining as much of the variation present as possible. PCA can be applied to dynamic structural response data to identify the predominant modes of vibration of the structure. Because PCA is a statistical technique, there are errors in the computed modes due to the use of a sample of finite size. The aim of this paper is to study the effect of sample size on the accuracy with which the modes of vibration can be computed. The paper focuses predominantly on elastic response data and examines the potential influence of various parameters such as the period of the structure, the input excitation, and the spatial distribution of mass over the structure. Issues relating to errors in the modes of nonlinear structures are also discussed.     

基于自适应梯度先验的旋转模糊图像复原算法

弹载红外成像系统随弹体作高速旋转运动时,在曝光时间内获取的图像存在严重旋转模糊问题,给后续的目标识别与图像跟踪造成极大困扰。针对此问题,提出了一种基于自适应梯度先验的旋转模糊图像复原算法。该算法通过使用自适应梯度先验的正则化项,对从图像中沿旋转模糊路径提取的一维向量在频域内进行反卷积运算。同时针对使用Bresenham算法提取像素方法产生的空穴点问题,设计了一种查表决策的自适应中值滤波算法。仿真实验结果表明,相对于改进的维纳滤波、约束最小二乘滤波、梯度加载滤波,该算法能有效地适应低信噪比干扰环境,具有较强的噪声抑制和削弱振铃效应的能力。     

A splitting-based finite element method for the Biot poroelasticity system

In this work, we propose a finite element method for solving the linear poroelasticity equations. Both displacement and pressure are approximated by continuous piecewise polynomials. The proposed method is sequential, leading to decoupled smaller linear systems compared to the systems resulting from a fully implicit finite element approach. A priori error estimates are derived. Numerical results validate the theoretical convergence rates.     

Fuzzy Numbers in Cost Range Estimating

Range estimating is a simple form of simulating a project estimate by breaking the project into work packages and approximating the variables in each package using statistical distributions. This paper explores an alternate approach to range estimating that is grounded in fuzzy set theory. The approach addresses two shortcomings of Monte Carlo simulation. The first is related to the analytical difficulty associated with fitting statistical distributions to subjective data, and the second relates to the required number of simulation runs to establish a meaningful estimate of a given parameter at the end of the simulation. For applications in cost estimating, the paper demonstrates that comparable results to Monte Carlo simulation can be achieved using the fuzzy set theory approach. It presents a methodology for extracting fuzzy numbers from experts and processing the information in fuzzy range estimating analysis. It is of relevance to industry and practitioners as it provides an approach to range estimating that more closely resembles the way in which experts express themselves, making it practically easy to apply an approach.     

Bridge user cost estimation – a synthesis of existing methods and addressing the issues of multiple counting,workzones and traffic capacity limitation

Efforts to incorporate the concerns of bridge users in bridge investment evaluation are often stymied by lack of a comprehensive framework for assessing different user costs. There is a need to synthesise and update existing user cost estimation techniques so that the incorporation of user costs in bridge investment evaluation can be more consistent and streamlined. Secondly, a bridge detour may occur for more than one reason, thus there is a danger of multiple counting that could cause overestimation of user costs. Thirdly, user costs during bridge workzones have rarely been considered in the literature. To address these issues, this article presents a framework for comprehensive estimation of bridge user costs, an approach to address the multiple-counting problem, and a methodology for bridge workzone user cost estimation. Furthermore, the article develops a method to estimate bridge user delay cost due to traffic capacity limitation. The methodologies are demonstrated using a case study.     

Association rule mining using binary particle swarm optimization

In this paper, we developed a binary particle swarm optimization (BPSO) based association rule miner. Our BPSO based association rule miner generates the association rules from the transactional database by formulating a combinatorial global optimization problem, without specifying the minimum support and minimum confidence unlike the a priori algorithm. Our algorithm generates the best M rules from the given database, where M is a given number. The quality of the rule is measured by a fitness function defined as the product of support and confidence. The effectiveness of our algorithm is tested on a real life bank dataset from commercial bank in India and three transactional datasets viz. books database, food items dataset and dataset of the general store taken from literature. Based on the results, we infer that our algorithm can be used as an alternative to the a priori algorithm and the FP-growth algorithm.     

结合边缘特征先验引导的深度卷积显著性检测

针对当前基于深度学习的显著性检测算法缺少利用先验特征和边缘信息，且在复杂场景中难以检测出鲁棒性强的显著性区域的问题，提出了一种结合边缘特征，利用先验信息引导的全卷积神经网络显著性检测算法。该算法利用三种被经常用到的先验知识结合边缘信息形成先验图，通过注意力机制将提取的先验特征与深度特征有效融合，最终通过提出的循环卷积反馈优化策略迭代地学习改进显著性区域，从而产生更可靠的最终显著图预测。经过实验定性定量分析，对比证明了算法的可靠性。     

Combining Upscaling and Downscaling of Methane Emissions from Rice Fields: Methodologies and Preliminary Results

The uncertainty in the methane (CH₄) source strength of rice fields is among the highest of all sources in the global CH₄ budget. Methods to estimate the source strength of rice fields can be divided into two scaling categories: bottom-up (upscaling) and top-down (downscaling). A brief review of upscaling and downscaling methodologies is presented. The combination of upscaling and downscaling methodologies is proposed as a potential method to reduce the uncertainty in the regional CH₄ source strength of rice fields. Some preliminary results based on upscaling and downscaling are presented and the limitations of the approaches are discussed. The first case study focuses on upscaling by using a field-scale model in combination with spatial databases to calculate CH₄ emissions for the island of Java. The reliability of upscaling results is limited by the uncertainty in model input parameters such as soil properties and organic carbon management. Because controlling variables such as harvested rice area may change on relatively short time scales, a land use change model (CLUE) was used to quantify the potential land use changes on Java in the period 1994–2010. The predicted changes were evaluated using the CH₄ emission model. Temporal scaling by coupling land use change models and emission models is necessary to answer policy-related questions on future greenhouse gas emissions. In a downscaling case study, we investigate if inverse modeling can constrain the emissions from rice fields by testing a standard CH₄ from rice scenario and a low CH₄ from rice scenario (80 and 30 Tg CH₄ yr^–1, respectively). The results of this study are not yet conclusive; to obtain fine-resolution CH₄ emission estimates over the Southeast Asian continent, the monitoring network atmospheric mixturing ratios need to be extended and located closer to the continental sources.     

Recommending Personalized POIs from Location Based Social Network

Location based social networks (LBSNs) provide location specific data generated from smart phone into online social networks thus people can share their points of interest (POIs).POI collections are complex and can be influenced by various factors,such as user preferences,social relationships and geographical influence.Therefore,recommending new locations in LBSNs requires to take all these factors into consideration.However,one problem is how to determine optimal weights of influencing factors in an algorithm in which these factors are combined.The user similarity can be obtained from the user check-in data,or from the user friend information,or based on the different geographical influences on each user's check-in activities.In this paper,we propose an algorithm that calculates the user similarity based on check-in records and social relationships,using a proposed weighting function to adjust the weights of these two kinds of similarities based on the geographical distance between users.In addition,a non-parametric density estimation method is applied to predict the unique geographical influence on each user by getting the density probability plot of the distance between every pair of user's check-in locations.Experimental results,using foursquare datasets,have shown that comparisons between the proposed algorithm and the other five baseline recommendation algorithms in LBSNs demonstrate that our proposed algorithm is superior in accuracy and recall,furthermore solving the sparsity problem.     

A modified hidden Markov model

This paper considers two discrete time, finite state processes X$X$ and Y$Y$. In the usual hidden Markov model X$X$ modulates the values of Y$Y$. However, the values of Y$Y$ are then i.i.d. given X$X$. In this paper a new model is considered where the Markov chain X$X$ modulates the transition probabilities of the second, observed chain Y$Y$. This more realistically can represent problems arising in DNA sequencing. Algorithms for all related filters, smoothers and parameter estimations are derived. Versions of the Viterbi algorithms are obtained.