Eye-safe compact scanning LIDAR technology

A 1.5-μm eye-safe, 3-D scanning, and compact Mie LIght Detection and Ranging (LIDAR) is presented. The transmitter of the LIDAR is based on a KTA optical parameter oscillator (OPO) resonator. For detecting return signals, an InGaAs APD is used. The all solid-state OPO laser transmitter has the feature of small volume and lightweight, which allows a 165-lb compact eye-safe scanning LIDAR to be constructed. A system simulation using our own model was conducted to direct the system development. A method to solve the problem with small active area APD detectors was developed and described. The preliminary field-test measurement results indicated that the LIDAR has the capability to detect aerosols and clouds in lower atmospheres up to three dimensions.     

基于支持向量机的多类分类研究

现今流行的分类方法的重要基础是传统的统计学，前提是要有足够的样本，当样本数目有限时容易出现过学习的问题，导致分类效果不理想。引入支持向量机方法，它基于统计学习理论，采用了结构风险最小化原则代替经验风险最小化原则，较好的解决了小样本学习的问题；又由于采用了核函数思想，把非线性空间的问题转换到线性空间，降低了算法的复杂度。对其相关内容包括优化算法及多类分类问题的解决进行了研究，最后用一个实例说明了该方法的可行性和有效性。     

基于支持向量机的缓存预估模型的设计与实现

采用支持向量机(SVM)对网络业务流数据进行预测估计，通过训练样本，从而获得样本以外数据的分布规律。在此基础上，设计了一种网络排队队列缓存的预估模型。实验表明，该模型具有较高的训练效率和很高的估计精确度。     

Pruning Redundant Alarm Correlation Patterns

1　ＩｎｔｒｏｄｕｃｔｉｏｎＴｈｅｓｔｉｃｋｉｎｇｐｏｉｎｔｏｆａｓｓｏｃｉａｔｉｏｎｒｕｌｅａｌｇｏｒｉｔｈｍｓｄｅｐｅｎｄｓｏｎｅｆｆｅｃｔｉｖｅｌｙｆｉｎｄｉｎｇａｌｌｃｏｒｒｅｌａｔｉｏｎｐａｔｔｅｒｎｓｔｈａｔｓａｔｉｓｆｙｖａｌｕｅｒｅｑｕｉｒｅｍｅｎｔｉｎｔｈｅｍａｇｎａｎｉｍｉｔｙｏｆｄａｔａ.Ｂｕｔｔｈｅａｌｇｏｒｉｔｈｍｓａｌｓｏｂｒｉｎｇａｎｅｇａｔｉｖｅｅｆ ｆｅｃｔ:ｔｈｅｎｕｍｂｅｒｏｆａｓｓｏｃｉａｔｉｏｎｒｕｌｅｓｉｓｖｅｒｙｌａｒｇｅ.Ａｌｓｏｉｎｆｏｒｍａｔｉｏｎ…     

On many‐sorted algebraic closure operators

A theorem of Birkhoff‐Frink asserts that every algebraic closure operator on an ordinary set arises, from some algebraic structure on the set, as the corresponding generated subalgebra operator. However, for many‐sorted sets, i.e., indexed families of sets, such a theorem is not longer true without qualification. We characterize the corresponding many‐sorted closure operators as precisely the uniform algebraic operators. (© 2004 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

The effect of voice lessons on the clinical and perceptual skills of graduate students in speech-language pathology

Two groups of 10 speech-language pathology graduate students were each given 7 weeks of singing lessons to determine whether voice lessons could have an effect on their clinical and perceptual skills. Pre-, mid-, and posttests to measure various skills were designed and implemented. With use of paired sample statistical testing, statistically significant results were obtained. In addition, the subjective responses of the students show that the lessons were effective in improving pitch perception, breath control, and legato production or easy onset. This study supports efforts to integrate curricula in vocal performance and speech-language pathology.     

Estimates for transition probabilities on a compact manifold

The purpose of this note is to describe a procedure for transferring familiar estimates for transition probabilities on RN to transition probabilities on compact manifolds.     

Compact directed acyclic word graphs for a sliding window

Suffix trees are a well-known and widely-studied data structure highly useful for string matching. The suffix tree of a string w can be constructed in O(n) time and space, where n denotes the length of w. Larsson achieved an efficient algorithm to maintain suffix trees for a sliding window. It contributes to prediction by partial matching (PPM) style statistical data compression scheme. Compact directed acyclic word graphs (CDAWGs) are a more space-economical data structure for indexing strings. In this paper we propose a linear-time algorithm to maintain CDAWGs for a sliding window.     

Second Order Nonlinear Evolution Inclusions Ⅰ： Existence and Relaxation Results

This is the first part of a work on second order nonlinear, nonmonotone evolution inclusions defined in the framework of an evolution triple of spaces and with a multivalued nonlinearity depending on both x（t） and x（t）. In this first part we prove existence and relaxation theorems. We consider the case of an usc, convex valued nonlinearity and we show that for this problem the solution set is nonempty and compact in C^1 （T, H）. Also we examine the Isc, nonconvex case and again we prove the existence of solutions. In addition we establish the existence of extremal solutions and by strengthening our hypotheses, we show that the extremal solutions are dense in C^1 （T, H） to the solutions of the original convex problem （strong relaxation）. An example of a nonlinear hyperbolic optimal control problem is also discussed.     

混合效应模型中的D-最优设计

本研究的是混合效应-Ⅱ模型，首先，给出二阶模型存在D-最优回归设计点的必要条件。其次，组合D-最优回归设计和D-最优区组设计BIBD构造D-最优设计并给出寻找新的D-最优设计点的方法。以两种有代表性的组合误差分布为例，阐明新的D-最优设计点的获得过程，并给出了数值结果。