基于用户行为感知的数字期刊服务推送研究

【目的/意义】数字期刊阅读在移动网络的推动下已成为互联网使用群体必不可少的活动,分析数字期刊 服务中的用户行为能够掌握用户阅读倾向,提高推送精确度。从用户行为感知的角度出发根据用户信息分析用户 行为轨迹,搭建推送模型,结合推荐算法和用户行为预测模型,实现数字期刊的用户个性化推荐。【方法/过程】本文 使用聚类算法和加权二部图算法计算用户相似度、寻找用户邻居簇,进而预测用户对数字期刊的评分,基于加权 Markov模型预测用户模型,生成数字期刊推送结果。【结果/结论】利用用户行为感知技术对数字期刊服务推送提 出了基本的研究理论框架,依照推荐算法设计对不同用户进行个性化推荐,增强用户信赖感,提高用户对数字期刊 阅读的满意度。     

网络电子信息搜索目标预测算法研究

随着互联网络应用的日益普及,网络电子信息如火如荼,如何为用户提供个性化的、一对一的服务是网上预览向深层次发展需要解决的问题.提出了根据用户测览记录计算用户兴趣强度的算法,实现了对用户阅览目标的预测.实验结果表明,该算法不但能够如实的反映出用户访问节点的频度,而且也可以模拟用户的兴趣随时间迁移的情况,使预测结果更加准确.     

基于有向带权图的Web用户浏览行为模型

建立反映用户浏览特征的Web用户浏览行为模型是Web挖掘的关键。本文通过有向带权图来描述用户浏览行为,首先通过数据集构造出了Web站点的拓扑结构,再引入Markov链模型来描述团体成员的网页访问活动,对用户浏览行为进行分析,并对链接的权值进行了计算,最后综合利用构造出的Web站点的拓扑结构和Markov链转移概率矩阵描述的用户浏览特征得到了描述用户浏览行为的有向带权图模型。     

基于Markov链的B2C网站频繁访问路径知识发现研究

对B2C网站的用户浏览行为进行了系统的研究。通过建立基于Markov链的动态模型,对用户浏览行为进行了深入的分析。进一步综合利用Web日志文件信息和Web站点的拓扑结构,对用户有关某页面的兴趣度进行了计算。提出了相关页面的聚类矩阵,对页面进行相关性聚类;对于相关的页面集合进一步挖掘,可以发现网站的频繁访问路径。     

基于最大熵功率谱估计的Hadoop高速数据访问

提出一种基于最大熵功率谱估计的Hadoop云平台下网络音视频数据特征挖掘方法,实现对数据信息的高速访问。构建数据挖掘Hadoop云平台和数据挖掘访问模型,设计最大熵功率谱特征提取算法,采用分段思想将同一时间段的视音频数据进行群体分割,分段提取最大熵功率谱特征。将提取的特征信息进行维度匹配分箱和溯源处理,实现信息恢复,最终完成高速数据访问。仿真测试表明,该算法能有效地实现对网络音视频数据的特征挖掘,提高访问效率,访问响应时间较当前方法缩短明显。     

江西城镇老年公寓的需求分析预测

近年来,人口老龄化问题越来越突出.加强老年公寓的建设,是解决传统家庭尤其是独生子女家庭父母养老的发展趋势.同时,加强老年公寓的筹划与建设,也有利于保持经济稳定增长、缓解社会就业压力.本文通过Markov链模型和年龄移算法等方法测算老年人口比例,对我省城镇老年公寓的需求进行分析预测,并提出了相应的发展目标建议.     

拟合用户偏好的个性化搜索

文章从用户偏好的角度对个性化搜索进行了优化研究,提出了基于语义关联树的查询扩展算法以及基于该算法的拟合用户偏好的个性化搜索系统架构。语义关联树可以灵活有效地控制查询扩展模型,在此之上的拟合用户偏好的个性化搜索系统具有用户偏好自学习能力。实验证明,该方法能有效提高文本检索的准确率。     

数据库集群负载均衡优化的Markov模型研究

针对数据库集群负载过重问题,本文提出了一种基于Markov的数据库集群负载均衡优化模型。该模型引入了基于两级事务管理器的中间件模型,提高写数据库操作的效率,并采用Markov模型预测采样周期内执行节点负载信息,并结合滑动窗口方法来更新管理节点内各执行节点的负载信息记录,然后针对综合性能指标对其进行负载均衡优化。实验表明,基于Markov的数据库集群负载均衡优化模型在可靠性和性能上得到了提高,该模型切实有效。     

一种基于模式的兴趣挖掘算法

提出一种基于模式的兴趣挖掘算法,通过查询日志获取访问序列,使用本体中的概念描述用户兴趣,提出一种计算兴趣得分公式,并根据兴趣得分将用户兴趣序列划分为不同的兴趣模式。本文算法实验结果显示,算法能够有效地实现用户兴趣挖掘。     

基于关联的聚类分析在个性化服务中的应用

运用关联规则,确定用户频繁访问模式,使用页面相似聚类分析对用户频繁访问集分类,挖掘具有相似访问兴趣的网络用户的浏览模式,并对相关算法作了改进,从满足用户个性化信息服务出发,给出一种基于关联的用户访问模式聚类方法.实践证明,将基于关联的聚类分析方法应用到个性信息服务中是有效的.     

国有商业银行效率变化及趋势分析

本文采用基于投入的数据包络分析(DEA)方法,研究我国国有商业银行在1994-2001年间的技术效率及其变化。我们将技术效率分解为纯技术效率和规模效率,识别技术有效和无效的银行,并确定规模报酬变化。根据技术效率平均估计来研究国有商业银行效率的总体变化趋势,推断出在亚洲金融危机期间国有商业银行的技术效率是很低的,这与当时情况相符合;同时还预测了国有商业银行效率发展趋势,并分析其原因。     

Beyond direct neighbourhood effects: higher-order interactions improve modelling and predicting tree survival and growth

It is known that biotic interactions are the key to species coexistence and maintenance of species diversity. Traditional studies focus overwhelmingly on pairwise interactions between organisms, ignoring complex higher-order interactions (HOIs). In this study, we present a novel method of calculating individual-level HOIs for trees, and use this method to test the importance of size- and distance-dependent individual-level HOIs to tree performance in a 25-ha temperate forest dynamic plot. We found that full HOI-inclusive models improved our ability to model and predict the survival and growth of trees, providing empirical evidence that HOIs strongly influence tree performance in this temperate forest. Specifically, assessed HOIs mitigate the competitive direct effects of neighbours on survival and growth of focal trees. Our study lays a foundation for future investigations of the prevalence and relative importance of HOIs in global forests and their impact on species diversity.     

联合树算法的研究与改进

贝叶斯网络是人工智能领域研究不确定知识的重要工具，联合树算法是一种应用广泛的贝叶斯网络推理算法。从基于消息传递的图模型、消息传递方案、计算效率、存储效率四个方面对联合树的三种算法进行比较。并对Hugin算法进行改进，通过具体贝叶斯网络的验证，改进算法能显著提高计算效率。     

两种数字图书馆发展趋势预测模型的比较

本文运用灰色系统中的GM(1，1)预测模型与逻辑斯谛曲线(Logistic Curve)预测模型，对近年来数字图书馆研究文献量的统计数据进行拟合预测，并结合实例对两种预测模型的预测结果进行对比分析，确定出了逻辑斯谛曲线是适合数字图书馆研究的预测模型。     

基于分类回归树和神经网络模型的零售商品促销数据分析与探讨

随着信息技术的不断发展,应用商业智能技术进行数据挖掘与分析对商家来说也越来越重要,分类回归树和神经网络算法是数据挖掘的经典算法,其广泛运用在数据分析、预测和评估等方面。文章分别运用分类回归树和神经网络算法对零售商品采取促销方案后收入变化的数据进行分析,并建立相应的模型对促销方案效果进行预测。     

Using Nyquist or Nyquist-like plot to predict three typical instabilities in DC–DC converters

By transforming an exact stability condition, a new Nyquist-like plot is proposed to predict occurrences of three typical instabilities in DC–DC converters. The three instabilities are saddle-node bifurcation (coexistence of multiple solutions), period-doubling bifurcation (subharmonic oscillation), and Neimark bifurcation (quasi-periodic oscillation). In a single plot, it accurately predicts whether an instability occurs and what type the instability is. The plot is equivalent to the Nyquist plot, and it is a useful design tool to avoid these instabilities. Four examples are used to illustrate the accuracy of this new plot to predict instabilities in the buck or boost converter with fixed or variable switching frequency.     

Novel machine learning technique for predicting teaching strategy effectiveness

In this paper, we present an approach for evaluating and predicting the student’s level of proficiency when using a certain teaching strategy. This problem remains a hot topic, especially nowadays when information technologies are highly integrated into the educational process. Such a problem is essential for those institutions that rely on e-learning strategies as various techniques for the same teaching activities and disciplines are now available online. In order to effectively predict the quality of this type of (electronic) educational process we suggest to use one of the well known machine learning techniques. In particular, a proposed approach relies on using logic circuits/networks for such prediction. Given an electronic service providing a teaching strategy, the mathematical model of logic circuits is used for evaluating the student’s level of proficiency. Given two (or more) logic circuits that predict the student’s educational proficiency using different electronic services (teaching strategies), we also propose a method for synthesizing the resulting logic circuit that predicts the effectiveness of the teaching process when two given strategies are combined. The proposed technique can be effectively used in the educational management when the best (online) teaching strategy should be chosen based on student’s goals, individual features, needs and preferences. As an example of the technique proposed in the paper, we consider an educational process of teaching foreign languages at one of Russian universities. Preliminary experimental results demonstrate the expected scalability and applicability of the proposed approach.     

城市公交客运量的组合预测研究

公交客运量预测是公交线网规划与改进的重要依据,以往学者多采用单一的预测方法进行预测,预测精度偏低。本文采用组合预测法对公交客运量进行预测,该方法提高了预测的精度,为科学地制定公交线网进行了有益尝试,带来了社会效益和经济效益的改变,与以往方法相比有了明显的增加。基于这样一种思路,本文以南昌市公交客运量为例,对南昌市公交客运量进行了模拟预测,并对预测结果进行了相关验算,得到了预期结果,显示出组合预测法比单一的预测方法优越,适用于城市公交客运量预测。     

K-d tree based approach for point location problem in explicit model predictive control

Explicit Model Predictive Control (EMPC) produces control laws defined over a set of polyhedral regions in the state space, and the online computation of EMPC is to find the corresponding control law according to a given state by searching in a lookup table, called point location problem. This paper presents an approach of constructing a hybrid data structure called constructed k-d tree(CKDT), which combines the k-dimensional tree (k-d tree) with the binary search tree (BST) for point location in such polyhedral sets. To maintain a ‘full’ and balanced constructed tree the number of affine control laws is used as the basis for choosing the candidate hyperplanes (HPs) during the main construction process of CKDT, thus increasing offline efficiency by reducing the number of candidate HPs requiring computation. This methodology can be applied to the EMPC of high dimensional problems as the k-d tree – a main part of the CKDT approach - has already been successfully used to solve high dimensional problems in the field of computer science and engineering. The method involves a trade-off between memory storage requirement and online efficiency. A complexity analysis of the approach in the runtime and storage requirements is provided. The advantages of the method are supported by two examples in the paper.     

基于纯技术效率指数的我国重点钢铁企业群划分研究

本文通过构建钢铁企业数据包络分析(DEA)模型对我国钢铁产业中15家具有代表性的钢铁企业技术效率进行测算,并剔除规模因素产生的影响,以纯技术效率指数为标准,对其进行集群划分,预测出到我国钢铁企业整体发展趋势,并对我国重点钢铁企业目前发展状况存在差距和改进方向进行深入分析。