大规模训练数据的支持向量机学习新方法

支持向量机不能直接对大规模的训练数据进行学习。提出一种新的减小样本集规模的方法；在映射后的高维空间中寻找两种类别的交界部分，交界部分上的样本作为学习样本。并且指出，不需涉及具体映射的形式只用核函数即可找到交界部分的样本。实验表明，新方法优于直接在低维样本空间中寻找交界部分样本的方法。     

Neuropsychological functioning of girls with attention-deficit/hyperactivity disorder followed prospectively into adolescence: Evidence for continuing deficits?

The current study prospectively followed girls with attention-deficit/hyperactivity disorder (ADHD), along with a matched comparison sample, 5 years after childhood neuropsychological assessments. Follow-up neuropsychological measures emphasized attentional skills, executive functions, and language abilities. Paralleling childhood findings, the childhood-diagnosed ADHD group displayed moderate to large deficits in executive/attentional performance as well as in rapid naming relative to the comparison group at follow up (Mage 14.2 years). ADHD-inattentive versus ADHD-combined contrasts were nonsignificant and of negligible effect size, even when a refined, sluggish cognitive tempo subgroup of the inattentive type was examined. Although ADHD versus comparison group differences largely withstood statistical control of baseline demographics and comorbidities, control of childhood IQ reduced executive function differences to nonsignificance. Yet when the subset of girls meeting diagnostic criteria for ADHD in adolescence was compared with the remainder of the participants, neuropsychological deficits emerged even with full statistical control. Overall, childhood ADHD in girls portends neuropsychological and executive deficits that persist for at least 5 years. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

关于Smarandache可乘函数的均值研究

本文的主要目的是利用初等方法研究Smarandache可乘函数在简单数集中的均值性质,并给出一个有趣的渐近公式。     

Analysis of an algorithm for solution of conditional optimization problems with linear-fractional objective functions over permutations

A method is considered to solve a conditional optimization problem with a linear-fractional objective function over permutations. The performance of sub algorithms to solve this problem is evaluated. The practical efficiency of the algorithm is analyzed by conducting numerical experiments. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 4, pp. 133–146, July–August 2007.     

人工冻土热工参数智能反演方法研究

首先运用遗传算法,在一定的初边值条件下通过对比试验实测值与解析解确定出导热系数与比热的函数值α(λ,c,ρ),再利用已确定出的系数α,通过遗传数值法计算值与实测值比较,进而确定出等效导热系数λ与比热c。     

Measuring consistency in TTL-based caches

We propose a general modeling framework to evaluate the performance of cache consistency algorithms. In addition to the usual hit rate, we introduce the hit^* rate as a consistency measure, which captures the fraction of non-stale downloads from the cache. We apply these ideas to the analysis of the fixed TTL consistency algorithm in the presence of network delays. The hit and hit^* rates are evaluated when requests and updates are modeled by renewal processes. Classical results on the renewal function lead to various bounds.     

A new stereological principle for test lines in three-dimensional space

A new principle is presented to generate isotropic uniform random (IUR) test lines hitting a geometric structure in three-dimensional space (3D). The principle therefore concerns the estimation of surface area, volume, membrane thickness, etc., of arbitrary structures with piecewise smooth boundary. The principle states that a point-sampled test line on an isotropic plane through a fixed point in 3D is effectively an invariant test line in 3D. Particular attention is devoted to the stereology of particles, where an alternative to the surfactor method is obtained to estimate surface area. An interesting case arises when the particle is convex. The methods are illustrated with synthetic examples.     

静态球对称时空中的标量场辐射

采用多项式函数代替乌龟坐标的方法,给出静态球对称时空中黑洞事件视界和宇宙视界区域内标量场方程的解,并对事件视界附近势垒的性质进行了讨论。发现对带电量不同的时空,标量场辐射的形式亦有所不同;势垒随宇宙学常数的减小而变窄、变高。     

Cycle detection using a stack

We present an algorithm for detecting periodicity in sequences produced by repeated application of a given function. Our algorithm uses logarithmic memory with high probability, runs in linear time, and is guaranteed to stop within the second loop through the cycle. We also present a partitioning technique that offers a time/memory tradeoff. Our algorithm is especially well suited for sequences where the cycle length is typically small compared to the length of the acyclic prefix.