基于领域知识的个性化推荐模型及其应用研究

首先,讨论了传统电子商务领域个性化推荐算法及其应用,提出传统推荐方法通常忽视了领域知识,因而在实际应用中存在局限性;然后,讨论了领域知识的相关概念和分类,并在此基础上建立了结合领域知识的两阶段推荐模型;最后,基于网络超市数据进行了实证研究,实验结果表明改进后的推荐算法效果要明显优于传统的个性化推荐算法.     

电子商务推荐系统中群体用户推荐问题研究

尽管传统的电子商务推荐系统在个体用户推荐方面取得了巨大成功,但它并不适用于向群体用户进行推荐。随着虚拟社区中群体用户的不断增加,构建群体推荐系统,向群体用户提供个性化推荐,减少他们搜集信息所耗费的时间和精力显得越来越重要。基于此,本文提出了一种新颖的推荐方法—结合领域专家法的群体用户推荐算法。该算法以基于项目的协同过滤技术为基础,根据群体成员间的相互作用确定群体偏好,由群体偏好产生推荐,推荐过程中存在的成员未评分项采用领域专家法进行预测填充,此外本文算法还考虑了成员间相似关系对推荐质量的影响。实验结果表明了本文算法的有效性。     

基于改进广义S变换的井筒波压制

VSP资料中经常出现能量比较强的井筒波,目前其压制方法主要利用中值滤波和F-K滤波。由于中值滤波存在统计效应,而F-K变换经常存在空间假频,都很难取得理想效果。基于井筒波和有效信号的时频差异,提出利用改进的广义S变换(采用宽度可变的高斯窗函数,其时窗宽度随频率呈正比,在低频段时窗较窄,获得较高的时间分辨率;高频段时窗较宽,故可获得很高的频率分辨率)对井筒波进行分离和压制。理论推导和实际数据计算表明,基于吸收改进的广义S变换的井筒波压制明显突出了有效信号,提高了信噪比。      

The Ornstein-Uhlenbeck Dirichlet process and other time-varying processes for Bayesian nonparametric inference

This paper introduces a new class of time-varying, measure-valued stochastic processes for Bayesian nonparametric inference. The class of priors is constructed by normalising a stochastic process derived from non-Gaussian Ornstein-Uhlenbeck processes and generalises the class of normalised random measures with independent increments from static problems. Some properties of the normalised measure are investigated. A particle filter and MCMC schemes are described for inference. The methods are applied to an example in the modelling of financial data.     

基于多相滤波的宽带数字化接收机技术

提出了一种多相滤波的宽带数字化接收机的设计技术;讨论了多相滤波的基本原理、设计及FPGA实现,给出了多相滤波的中频数字化接收机的实现方案。计算机仿真结果和现场可编程门阵列仿真结果表明,多相滤波的中频数字化接收机设计方法是可行的,并还可灵活地实现现有专用芯片很难实现的宽带接收技术。     

一种新的自适应卡尔曼滤波算法——ALR算法

本文提出一种新的自适应卡尔曼滤波算法——ALR(Adaptive Least Residual)算法.该算法从经典的卡尔曼滤波出发,使残差达到最小,以残差(residual)样本值代替统计值,再用最陡下降法估计算自适应权增益,最后得到一组自适应滤波公式.通过理论分析和计算机模拟,证明该算法具有算法简单、运算量小等优点,适于在自适应系统辨识、自适应谱估计中作实时处理.     

The Effects of Publication Selection on Test Probabilities and Estimator Distributions

Editorial decisions, based in part on reported hypothesis test results, affect the probabilities associated with those results: the probabilities of Type I and Type II errors thus become different for readers than for authors. The distributions of published parameter estimates are similarly affected. A framework for studying the consequences of test-based information filtering is developed and illustrative examples are provided. The examples indicate that filtering can markedly distort the power functions of hypothesis tests and can induce large estimator biases and increases in mean square error. It is argued that test-based filtering is relevant not only to journal publication but to other forms of information dissemination as well.     

How flat organizations filter: organizational gatekeeping in a networked environment

The development of affordable Internet-based tools and platforms has led to major transformations in organizational dynamics. One key change is the shift of agency from center to ends and the consequent emergence of decentralized (networked) organizations. Focusing on the realm of collective action, we suggest that organizational configurations constructed around participants with greater agency induce organizations to promote new filtering tactics to sort out members who are valuable to the group. Drawing on Signaling Theory, we analyze the filtering behavior that organizations develop in their attempt to cope effectively with a type of tension characteristic of the digital era: The impetus to uphold egalitarian ideals on the one hand and the need to sort out desired participants on the other. Previous studies have explained how boundary work actually works in bureaucratic settings with fairly clear boundaries. This study focused instead on how new patterns of collective participation emerge, and how novel practices related to filtering are distributed in decentralized organizations. The study makes two points with theoretical implications. First, it explains an organization's decision to conduct filtering measures – a decision that on the face of it is counterintuitive since organizations operating in a digital environment can tolerate the enclosure of ‘free riders’. The second point relates to filtering practices. It shows that the emergence of new forms of participation, practices and norm, due to the development of networked media, encourages creativity in the development of specific filtering tactics, with activists considering new tactics for achieving their goals.     

西方居住隔离理论:发展历程与现实启示

自恩格斯首次揭示不同群体的居住隔离状态以来,西方社会学经济学等相关学科先后形成了三种不同取向的居住隔离理论,即强调不同人群空间分离的人文区位学居住分异理论,强调住宅更替规律的住房过滤理论,以及强调住宅阶级的新都市社会学居住隔离理论。这些理论共同提醒人们,在当代中国的城镇发展和住房建设中,必须有效预防居住空间隔离,阻止不同社会群体和阶层之间的隔离和对抗。     

Basil Bernstein: Theory,models and the question of method

The purpose of this article is to address the interpretive synthesis of qualitative research. The article discusses the potential to develop a ‘meta-interpretation’ approach which attempts to maintain an interpretive epistemology congruent with the majority of primary qualitative research. The article identifies and discusses a range of issues in the synthesis of interpretive qualitative research (objectivity, evaluation of studies, sampling, context, interpretation and integrity) that emerge from a review of nine research methods involving synthesis and a broader review of the research synthesis literature. Seven fundamental features of meta-interpretation are outlined and the meta-interpretation procedure is presented. In conclu-sion, the aims of synthesis procedures in general and meta-interpretation in particular are discussed.