基于领域本体的网页主题相关度计算

为提高网页内容与特定主题之间相关度计算的准确度,提出一种基于领域本体的网页主题相关度计算模型OBWTCCM(ontology based webpage-topic correlation calculation model)。使用领域本体刻画主题,通过计算本体概念间的语义关系提取主题概念并构造主题语义矩阵,将特征词的统计信息与该矩阵相结合计算网页与主题之间的相关度。该模型改进了向量空间模型在相关度计算时对特征词语义层次分析的不足。实际项目应用结果表明,使用该方法计算得到的网页主题相关度与领域专家的判断总体相符,具有较理想的准确度。     

领域同义特征词的统计规律及其在情感分析上的应用研究

为了研究同义特征词的规律及其对情感分析的影响,基于大规模语料库的统计学习方法,发现如果将每一个特征词的所有同义特征词组成特征词组,词组中的同义特征词按照自身在语料库中出现的频率从大到小排列,那么同义特征词在词组中出现的频率与该词在词组中的名次存在幂函数关系：y＝β0xβ1。将特征代表词的同义特征词所对应的情感词加入到该特征代表词的情感词集中,可以有效提高特征代表词的情感分析的准确性。     

一种基于词聚类的中文文本主题抽取方法

提出了一种基于词聚类的中文文本主题抽取方法,该方法利用相关度对词的共现进行分析,建立词之间的语义关联,并生成代表某一主题概念的用种子词表示的词类。对于给定文档,先进行特征词抽取,再借助词类生成该文档的主题因子,最后按权重输出主题因子,作为文本的主题。实验结果表明,该方法具有较高的抽准率。     

一种新的微博短文本特征词选择算法

针对微博短文本有效特征较稀疏且难以提取,从而影响微博文本表示、分类与聚类准确性的问题,提出一种基于统计与语义信息相结合的微博短文本特征词选择算法。该算法基于词性组合匹配规则,根据词项的TF-IDF、词性与词长因子构造综合评估函数,结合词项与文本内容的语义相关度,对微博短文本进行特征词选择,以使挑选出来的特征词能准确表示微博短文本内容主题。将新的特征词选择算法与朴素贝叶斯分类算法相结合,对微博分类语料集进行实验,结果表明,相比其它的传统算法,新算法使得微博短文本分类准确率更高,表明该算法选取出来的特征词能够更准确地表示微博短文本内容主题。     

融合统计信息与语义相似度的特征扩展算法

通过分析短文本的高维性和稀疏性,提出一种融合特征词间统计信息与语义相似度的短文本特征扩展算法。根据词的贡献度对候选特征集进行筛选,得到扩展集合初始值。计算特征词之间的统计相关度,构建二元相关词对集合。利用外部知识库知网中的语义关系获取相关词对的义项集合并计算语义相似度,将满足条件的义项扩展为短文本的特征词,得到扩展后的特征集。实验结果表明,使用该算法对短文本进行特征扩展后,可显著提升分类器的分类效果。     

从高频词等级相关角度探析《红楼梦》作者

该文提出一种“基于高频词等级相关度的方法”来探析存疑文献的作者信息,把各份语料中的词型均按照出现频次递减排列并确定等级,然后通过计算出语料之间高频词等级的相关度,来推断语料之间语言风格的相似度,并且把这种方法与“基于词型共现率的方法”和“基于词例共现率的方法”相比较。把《红楼梦》的120回均分为12份语料,使用“基于高频词等级相关度的方法”计算这12份语料两两之间的相关度。研究发现《红楼梦》的前8份语料两两之间相关度高,后4份语料两两之间相关度也高,而前8份语料与后4份语料这两部分语料之间相关度低。推断《红楼梦》前80回应是同一人所写,后40回应是另一人所写。     

基于语义的文档特征提取研究方法

中文文本特征词选取是文本处理的重要方面,对文本分类有重要影响。现有的文本特征提取方法存在生成特征向量维数高、依赖训练集、忽略低频关键词等不足。利用《同义词词林》计算词语之间的语义距离,通过聚类算法筛选类别的主题相关词,最后通过信息增益算法从主题相关词中选取特征词。以宏F值和微F值为评价指标,通过有效性实验和对比实验表明,该方法的文本特征选取效果优于其他经典算法。     

主题搜索引擎中特征模型技术的研究

主题搜索引擎的研究难点之一就是主题与网页信息之间的准确匹配。通过对网页的特征进行分析,提取网页特征中的主题特征词,并用提取的主题特征词表示网页主题信息,提出了利用网页特征及特征之间的关系来建立网页特征模型。该特征模型能准确地描述网页的内部特征和外部特征的主题表现力,有利于计算网页与主题之间的相似度。实验结果表明该特征模型能有效地表达网页的主题信息,并有助于提高主题搜索引擎的资源发现率和搜索准确率。     

基于领域本体和位置关系的信息检索模型

向量空间模型是最常用的信息检索模型,它根据词频来计算文档之间的相关度,这种方法虽然能够满足用户的基本检索需求,但是对于检索要求较高的用户,其效果仍然不甚理想。文中在向量空间模型的基础上,首先通过领域本体和上层本体来计算特征词项之间的相似度,据此得出与查询词相关的词,在求词项频率和逆文档频率时考虑这些词,然后引入了词序相关度和词语相邻相关度这两个概念,把特征项的位置关系也考虑进来。实验结果表明,文中提出的模型相比原始向量空间模型,在准确率上有了较大的改善。这完全说明,与原始向量空间模型相比,文中提出的检索模型不仅考虑了与原有词项具有相似语义的词项,而且还考虑了词项顺序和词项相邻信息,从而更能符合用户的检索要求。     

基于DMA与特征划分的多源文本主题模型

针对传统主题模型在挖掘多源文本数据集信息时存在主题发现效果不佳的问题,设计一种基于狄利克雷多项式分配（DMA）与特征划分的多源文本主题模型。以DMA模型为基础,放宽对预先输入的主题数量的限制,为每个数据源分配专有的主题分布参数,使用Gibbs采样算法估计每个数据源的主题数量。同时,对每个数据源分配专有的噪音词分布参数以及主题-词分布参数,采用特征划分方法区分每个数据源的特征词和噪音词,并学习每个数据源的用词特征,避免噪音词集对模型聚类的干扰。实验结果表明,与传统主题模型相比,该模型能够保留每个数据源特有的词特征,具有更好的主题发现效果及鲁棒性。     

基于大数据的台区行业聚合分类方法及分类特征分析

在研究台区近中期负荷预测方法的过程中,遇到了如何利用大数据识别台区进行行业分类的问题。经过研究,将这个问题分为台区行业分类方法和行业负荷特征两方面。台区行业分类确定了以用电类别作为一级分类,以及运用数据挖掘中的k-means算法对台区典型日年（最大）负荷曲线进行聚类的二级分类共同组成的分类方法;行业负荷特征研究在台区行业分类的基础上,分析行业负荷特征,包括典型日负荷特征和年负荷特征。并以此方法在深圳大数据平台对深圳市台区进行行业分类和分类特征分析。行业分类中将公专变台区一级分类后,都居民生活台区进行聚类分析,分别形成以居民负荷和学校负荷为主的两类。行业负荷特征分析中以学校台区为例,以学生是否住宿为分别,可以区分出走读类学校和住宿类学校。结果表明,此方法效果良好。     

融合双重权重机制和图卷积神经网络的微博细粒度情感分类

利用深度学习模型和注意力机制对微博文本进行细粒度情感分类,已成为研究的热点,但是现有注意力机制只考虑单词对单词的影响,对单词本身的多种维度特性(如词义、词性、语义等特征信息)缺乏有效的融合.为了解决这个问题,文中提出了一种双重权重机制WDWM(Word and Dimension Weight Mechanism),并...     

Internet技术教学改革与创新

结合学生的学科专业特点和学校区域分布,对学生进行学科分类,同时对Internet技术进行层次划分,形成按学科分层次的Internet技术教学特点。在课程教学改革与创新过程中,采取多种形式增强学生Internet技术的应用能力。     

稀疏数据下结合词向量的短文本分类模型研究

针对短文本缺乏足够共现信息所产生的词与词之间弱连接,且难以获取主题词的情况,导致面向短文本分类工作需要人工标注大量的训练样本,以及产生特征稀疏和维度爆炸的问题,提出了一种基于注意力机制和标签图的单词共生短文本分类模型(WGA-BERT)。首先利用预先训练好的BERT模型计算上下文感知的文本表示,并使用WNTM对每个单词的潜在单词组分布进行建模,以获取主题扩展特征向量;其次提出了一种标签图构造方法捕获主题词的结构和相关性;最后,提出了一种注意力机制建立主题词之间,以及主题词和文本之间的联系,解决了数据稀疏性和主题文本异构性的问题。实验结果表明,WGA-BERT模型对于新闻评论类的短文本分类,比传统的机器学习模型在分类精度上平均提高了3%。     

Models of pedagogical implementation of ICT in Israeli schools

Abstract The paper analyses the data collected in 10 schools in Israel which have incorporated ICT in unique ways and have succeeded in devising innovative classroom pedagogies and changes in teachers and students roles and outcomes. The major research questions addressed are: What levels of innovation were observed in the participant schools, and in which domains? How did the level of innovation in the various domains vary among and within schools? What correlation patterns among levels of innovation in the various domains can be identified? The data were analysed by mean of the innovation analysis schema developed for characterising ICT‐based educational innovations. The findings indicate that most schools were in a transition stage towards fully innovative pedagogical implementation; school activities were effected differentially by ICT, thus creating ‘islands of innovation’; these islands however, have the potential to pull forward other areas of activity and people in the school; the domains most affected by ICT were the school's digital space, the didactic solutions implemented and the students' roles and outcomes.     

A multidimensional approach to determinants of computer use in primary education: teacher and school characteristics

The central aim of this study was to test a model that integrates determinants of educational computer use. In particular, the article examines teacher and school characteristics that are associated with different types of computer use by primary school teachers. A survey was set up, involving 527 teachers from 68 primary schools in Flanders. A separate questionnaire was administered to information and communication technology (ICT) coordinators from the same schools to gather additional information about cultural and contextual school characteristics. The combined impact of both teacher and school characteristics was explored through a multilevel analysis. Besides the importance of school characteristics, the results reveal differential effects of specific characteristics on specific types of computer use. Cultural school characteristics for instance, such as the schools’ openness to change and the availability of an ICT school policy plan, are positively related to the use of computers as a learning tool and to the adoption of ICT in view of basic computer skills. In contrast, no cultural school characteristic seem to be associated with the use of computers as an information tool. In a comparable way, teacher characteristics are associated with specific types of computer use, e.g. the variable gender. In general, male teachers report integrating computers more often. In this study, it appears that gender differences only exist in relation to the adoption of computers as an information tool. The results demonstrate that a multidimensional approach provides more insight into the characteristics affecting computer use.     

How trainee teachers of IT approach teaching their subject

Abstract  This paper explores the nature of Information Technology (IT) as a separate subject in the secondary school curriculum. It reports on research carried out with a cohort of specialist IT post-graduate trainee teachers at a local university. The paper explores IT trainees' approach to teaching their subject and describes their strong focus on purpose when teaching IT. Important characteristics of the purposeful approach are the use of authentic data; the transformation of data by pupils; and a focus on problem solving within a context to which pupils can relate. However in their work in schools, trainees found constraints on developing a strong focus on purpose in their lessons. They were encouraged to experiment but felt they needed more support to carry out purposeful activities with pupils. The implications of this research on teacher training and on the value of IT as a curriculum subject are discussed.     

校园网与教育城单点登录系统的异构对接

广州市教育部门搭建了“数字教育城”,可以将各学校的应用系统通过单点登录整合到其中,以实现较高级别 的资源共享。但由于各学校的实际情况差异较大,整合过程必然会存在一些客观问题。本文探讨了广州市中职、中小学在接 入广州市数字教育城单点登录系统中可能遇到的部分特殊问题,尤其对已经有自己的单点登录方案的学校如何进行异构对 接,提出了比较适合现阶段实际情况的良好解决方案。     

Effects of an intelligent web‐based English instruction system on students' academic performance

This research conducted quasi‐experiments in four middle schools to evaluate the long‐term effects of an intelligent web‐based English instruction system, Computer Simulation in Educational Communication (CSIEC), on students' academic attainment. The analysis of regular examination scores and vocabulary test validates the positive impact of CSIEC, and in most cases, the positive impact is statistically significant. The reliability is ensured by the spectrum of the students from Grade 1 to Grade 3 in three junior high schools and from Grade 1 to Grade 2 in one senior high school, and of the teachers with or without blended learning experience, as well as by the various school locations from rural to urban areas in four provinces of China. The learning content‐oriented design and the instant feedback feature of the web‐based system, as well as its regular integration into the English class, contributed to its reliable positive effect on students' learning performance in ordinary examinations. This is the research's implication for instructional design. The conclusion of this paper could serve as a reference to the technical feasibility and pedagogical benefit of regular usage of appropriate Computer Assisted Learning system in the mainstream subject of middle schools. The multiple quasi‐experiments in divergent school settings to ensure the results' reliability distinguish this research from previous ones that just implemented one single experiment in one school.     

现代教育技术与各学科课程整合的探讨

随着现代教育技术的日益发展和成熟,现代教育技术与各学科课程的整合也日益为人们所关注,它被越来越广泛的应用于学校的教学当中,在帮助教学活动的过程中,现代教育技术起着非常重要的作用。它将彻底改变传统教学的过程和模式。但是,我们不可能让每个学科的教师都能象教育技术专业的老师一样软件样样会用,对于老师来说最直接最实用的应该是理解整合的理念,将这种思想和自己学科的教学智慧地结合,解决实际问题,怎样整合才能适应中国教育的现状,获得最佳效果,我们将在实践中去不断的探索。该文将对整合的定义、教学的指导思想、主要技术特点等进行了详细的阐述。