一种基于图的情感基准词选择方法

作为文本情感分析的前提和基础,词语的情感极性判别显得尤为重要。现有利用情感基准词进行词语的情感倾向研究中,情感基准词的选择多数基于研究者的人工判别或词语的使用频率。以上方式存在着随机性和主观性的缺陷,并且难以保证对词典中语义关系的全面覆盖。本文提出以候选基准词为顶点,两词间的知网相似度作为边的权重设定参数来构建情感词的无向图。将图中结点的中介性值作为基准词的选择依据,从而保证所选基准词的可靠性。实验证明,通过该方法选取出来的基准词在词的情感倾向分类中具有较高的准确率。     

基于DC-Value的西班牙语文本词语提取方法

西班牙语(以下简称西语)是仅次于汉语的世界第二大母语语言, 是联合国6种官方语言之一. 西语复杂的词形变化和语法规则, 导致C-value等经典的词语提取方法的效果无法保证, 进而影响基于西语文本挖掘的效果.因此, 本文研究西语文本词语提取方法, 为西语文本的结构化建模提供完备的词库. 给定待分析的西班牙语文本, 该方法分3步提取得到词语集合: 文本预处理、候选词语提取和DC-value成词度计算. 其中, 前两步所得的候选词语集合可直接用作文本挖掘的词库; 第三步所得的候选词语成词度可辅助判断候选词语成词的可能性, 减轻人工判断的工作量. 实验结果表明, 本文方法自动提取的西文词语集合的准确率达到80％, 且召回率远高于经典方法, 能够为西语文本挖掘提供有效的词库.     

基于连词的维吾尔语情感词库扩展研究

极性情感词是准确分析维吾尔文倾向性的基础资源。该文在前期构建的维吾尔语褒贬情感词典基础上进行网络情感词的自动扩展研究。首先分析维吾尔语情感表达的语言特征,总结了连词、程度副词与情感词的搭配规律,并基于此规律设计从情感语料库中获取候选情感词的算法,形成候选情感词库;最后再利用维吾尔语连词的特性,结合已创建的情感词典和维吾尔语反义词词典,以互联网作为超大规模语料库,设计基于搜索引擎的情感词极性判别算法,根据算法得分判别候选情感词的极性,再将其扩展到已构建的褒贬情感词库。实验结果表明,与扩展前的情感词库相比,使用互联网文本语料扩展后的情感词库后进行维吾尔语句子倾向性测评的准确率和召回率均有明显提高。     

结合卷积神经网络和词语情感序列特征的中文情感分析

目前基于词嵌入的卷积神经网络文本分类方法已经在情感分析研究中取得了很好的效果。此类方法主要使用基于上下文的词嵌入特征,但在词嵌入过程中通常并未考虑词语本身的情感极性,同时此类方法往往缺乏对大量人工构建情感词典等资源的有效利用。针对这些问题,该文提出了一种结合情感词典和卷积神经网络的情感分类方法,利用情感词典中的词条对文本中的词语进行抽象表示,在此基础上利用卷积神经网络提取抽象词语的序列特征,并用于情感极性分类。该文提出的相关方法在中文倾向性分析评测COAE2014数据集上取得了比目前主流的卷积神经网络以及朴素贝叶斯支持向量机更好的性能。     

面向评论信息的跨领域词汇情感倾向判别方法

词汇的情感倾向判别对文本情感分类具有重要意义。已有方法多假设存在基准词,根据目标词与基准词的关联度来判别目标词的情感倾向。实际应用中,尤其是评论语料库中基准词往往存在情感歧义问题,从而影响判别结果的准确性。基于上述分析,面向给定语料库,提出一种基准词的提取和消歧方法,并在此基础上实现跨领域的词汇情感倾向判别。首先在任一标记语料库中自动提取候选基准词;然后基于共现矩阵评估并过滤部分具有情感歧义的基准词;最后通过计算基准词与目标词的相似性,实现目标词的情感倾向判别。实验结果表明了方法的有效性和可行性。     

基于《知网》概念定义的情感词典构建研究

情感倾向,就是人对人或事物的看法,即主观色彩,通常分为褒贬、积极消极、好坏等方面。情感词语的情感倾向判别和权值赋予问题是文本倾向性分析研究中的基础,情感权值的研究在文本倾向性分析、舆情分析、文本分类等研究领域有着广泛的应用,最具有代表性的方法是通过对《知网》中词语的义原相似度的计算来进行词语相似度的计算。在其词语相似度计算方法的基础上,对《知网》词语概念库glossary.dat文件进行提取、修剪和增删,并通过同义词、反义词和人工甄选种子词语,使其对于情感词的权值的计算研究更加精确,实验结果表明,该方法在情感词褒贬义判别、权值取值上和应用上都有不错的效果。     

基于概念的中文博客情感极性聚类分析

通过基于概念的聚类方法,对博客作者的情感极性进行分析。在知网情感词汇库的基础上,将概念引入向量空间模型。首先,提取博客文本情感词,利用基于情感词概念的向量空间模型完成对博客文本的表示。然后,使用k-means算法对博客文本进行聚类,完成对博客情感极性的分析。在向量空间模型中使用概念作为特征项,提高了对博客作者情感极性分析的精度。实验证明基于概念的向量空间模型比传统基于词语的向量空间模型在博客文本情感聚类上具有更好的性能。     

基于注意力机制的评论情感分析及情感词检测

评论情感分析是用户生成内容分析的一个研究热点。评论对象的多样性与评论者用语的随意性,导致评论情感分析成为一个非常具有挑战性的任务。现有方法主要通过预先构建情感词表来计算评论的情感极性,但这类方法无法处理同一个词语在不同语境下情感极性存在差异的问题。针对这一问题,文中提出了一种基于注意力的卷积-递归神经网络模型,对评论的情感极性和词语在不同语境下的情感极性进行了建模。通过结合词语在句子中的上下文语境,所提方法可以将注意力集中在主要情感词周围的一个小范围内,并以一种自适应的方式对情感词的情感极性进行计算,提高了词语情感极性判断的准确率,进而提高了短文本的情感极性准确率。与CRNN,CNN以及基于情感词典的方法相比,所提方法在中文数据集(美团评论、党建评论)和英文数据集(亚马逊商品评论数据集)上都达到了更好的效果。     

基于卷积注意力机制的情感分类方法

为提取文本的局部最优情感极性、捕捉文本情感极性转移的语义信息,提出一种基于卷积注意力机制的神经网络模型(CNN_attention_LSTM)。使用卷积操作提取文本注意力信号,将其加权融合到Word-Embedding文本分布式表示矩阵中,突出文本关注重点的情感词与转折词,使用长短记忆网络LSTM来捕捉文本前后情感语义关系,采用softmax线性函数实现情感分类。在4个数据集上进行的实验结果表明,在具有情感转折词的文本中,该模型能够更精准捕捉文本情感倾向,提高分类精度。     

针对文本情感转换的SMRFGAN模型

文本情感转换的任务需要调整文本的情感并保留与情感无关的内容。但是由于缺乏并行数据,很难提取独立于情感的内容并以无监督学习的方式对情感进行转换,并且由于GAN处理文本类的离散数据效果不如处理连续数据,为此使用了强化学习（Reinforcement Learning）的方法来解决GAN处理离散数据的问题。强化学习的奖励机制来自完整序列上的GAN的判别器,并且用蒙特卡罗搜索方法对生成器进行优化,从而提高生成文本的准确性。为了将源文本中的情感词的极性进行转换,在长短记忆神经网络（LSTM）中增加了自注意力机制（self-attention）,再通过情感记忆模块（sentiment-memory）结合上下文来生成情感词极性反转后的文本作为SMRFGAN（Self-attention Memory Reinforcement learning GAN）预训练的真实数据。实验结果表明,该模型较好地解决了独立于情感内容进行情感转换的问题,BLEU评分有较好的提升。     

多语种网络文本快速新词抽取

从网络文本中提取新词是网络信息处理中的一个重要问题,在信息检索、文本挖掘、词典编纂、中文分词等领域中都有重要应用。本文提出了一种与语言无关的快速新词提取算法,首先针对后缀树的数据结构将多语言文本进行统一编码,然后使用改进的统计方法在双后缀树上以线性时间统计重复串与邻接类别,并计算字符串的整体度,同时通过剪枝大幅度减少计算量,在中、英文语料上较好地实现了新词的抽取及排序。     

网络意见挖掘、摘要与检索研究综述

网络上带有人的主观感情色彩的评论性文本反映了人们的意见、态度和立场,因而具有很大的利用价值.信息挖掘技术针对这些主观文本进行处理,获得有用的意见、结论和知识.首先介绍了意见挖掘出现的背景和应用意义,然后从词汇情感极性识别、粗粒度的情感分类、细粒度的意见挖掘与摘要、意见检索和相关语言资源与系统5个方面综述了研究历程和现状,最后总结了研究难点与研究趋势.     

基于直推式学习的中文情感词极性判别

态度挖掘是近年来文本挖掘领域的热点课题之一,旨在发现文本中作者的主观态度倾向,为基于舆情的决策过程提供支持。目前已有的态度挖掘算法绝大多数都基于情感词典来识别情感词,在此基础上判别句子或文本的总体态度倾向。然而,手工构造和维护一部完善的情感词典是不现实的。对中文情感词的极性判别问题进行了研究,提出了基于直推式学习的中文情感词极性判别算法。该算法以少量情感词为种子,利用词典中词汇的解释信息,直推出其他词的情感极性。与使用相同情感种子词的解释信息作为训练数据的有监督学习算法相比,直推式学习算法的识别精度提高了20%左右。     

Generating American Sign Language animation: overcoming misconceptions and technical challenges

Misconceptions about the English literacy rates of deaf Americans, the linguistic structure of American Sign Language (ASL), and the suitability of traditional machine translation (MT) technology to ASL have slowed the development of English-to-ASL MT systems for use in accessibility applications. This article traces the progress of a new English-to-ASL MT project targeted to translating texts important for literacy and user-interface applications. These texts include ASL phenomena called “classifier predicates.” Challenges in producing classifier predicates, novel solutions to these challenges, and applications of this technology to the design of user-interfaces accessible to deaf users will be discussed.     

机器翻译系统文法中语法和语义研究——《天语》英汉翻译系统的文法概要

机器翻译是双语言背景下的言语行为,是双语言知识的计算机处理。一个机器翻译系统的性能优劣,取决于它所使用的文法(当然,算法对于系统的影响也不可忽视)对相关语言知识,特别是对语法、语义的描写的深度和广度。本文简要介绍了《天语》英汉机器翻译系统(TYECT)的文法的基本内容。     

A procedure for the evaluation and improvement of an MT system by the end-user

Evaluation of MT systems can be performed by developers or by end-users. In the following paper, we present an evaluation procedure (-test) which involves both. It will be used to evaluate the ARIANE French to English system.The quality of the MT system will be evaluated in terms of efficiency and up-grading capacity. Several thousands of pages will be translated by the system and revised by professional translators. This will give us the necessary figures to measure efficiency through cost saving. Problem sheets, filled in by the revisors, will be sent to the developers, who will use them to improve the MT system. Up-grading capacity will be evaluated with the help of these problem sheets and also by re-translating some parts of texts, at the end of the test, and comparing both versions.     

Promoting contextual vocabulary learning through an adaptive computer‐assisted EFL reading system

The study developed an adaptive computer‐assisted reading system and investigated its effect on promoting English as a foreign language learner‐readers' contextual vocabulary learning performance. Seventy Taiwanese college students were assigned to two reading groups. Participants in the customised reading group read online English texts, each of which was customised by the developed system to offer immediate and repeated meetings with previously encountered unknown words, while participants in the typical reading group read online texts without control on unknown word recurrence. After the 4‐week online reading treatment, the two groups of learners were tested by online immediate and delayed vocabulary tests, and the students from the customised reading group were also required to complete the questionnaire regarding system use. The results showed that the vocabulary mean scores from both immediate and delayed testing demonstrated significantly better results in word gain and word retention with the customised reading group and that the adaptive reading system was appealing for the students. The study demonstrated that by providing customised reading with word recurrence specific to individuals' unknown words, this adaptive computer‐assisted English as a foreign language reading system creates a more favourable condition for foreign language vocabulary growth.     

Opinion mining from noisy text data

The proliferation of Internet has not only led to the generation of huge volumes of unstructured information in the form of web documents, but a large amount of text is also generated in the form of emails, blogs, and feedbacks, etc. The data generated from online communication acts as potential gold mines for discovering knowledge, particularly for market researchers. Text analytics has matured and is being successfully employed to mine important information from unstructured text documents. The chief bottleneck for designing text mining systems for handling blogs arise from the fact that online communication text data are often noisy. These texts are informally written. They suffer from spelling mistakes, grammatical errors, improper punctuation and irrational capitalization. This paper focuses on opinion extraction from noisy text data. It is aimed at extracting and consolidating opinions of customers from blogs and feedbacks, at multiple levels of granularity. We have proposed a framework in which these texts are first cleaned using domain knowledge and then subjected to mining. Ours is a semi-automated approach, in which the system aids in the process of knowledge assimilation for knowledge-base building and also performs the analytics. Domain experts ratify the knowledge base and also provide training samples for the system to automatically gather more instances for ratification. The system identifies opinion expressions as phrases containing opinion words, opinionated features and also opinion modifiers. These expressions are categorized as positive or negative with membership values varying from zero to one. Opinion expressions are identified and categorized using localized linguistic techniques. Opinions can be aggregated at any desired level of specificity i.e. feature level or product level, user level or site level, etc. We have developed a system based on this approach, which provides the user with a platform to analyze opinion expressions crawled from a set of pre-defined blogs.     

面向中文文本的情感信息抽取语料库构建

情感信息抽取是情感分析中的一个重要子任务。虽然该任务已经开展有一段时间,但是面向中文文本的情感信息抽取任务研究才刚刚起步。目前中文文本的情感信息抽取面临的首要困难在于现有的相关中文语料库还非常有限。为了更好开展中文文本的情感信息抽取研究,该文重点研究了中文语料标注体系,构建一个规模较大、标注类型丰富的中文情感信息抽取语料库。除了常见语料库标注的情感倾向性、评价对象、情感词等信息外,重点标注了评价对象的省略、无情感词情感句表达及极性转移等情况。由语料信息统计可知,该文所指出的特殊现象(例如,评价对象的省略)在中文情感表达中是非常普遍的,开展这方面的研究很有必要。该文所构建的中文文本语料库将为中文情感信息抽取任务提供语料基础。     

Problems of machine translation of business texts from Russian into English

This article draws on the example of business texts to consider practical aspects of the distortion of meaning in translation from one language to another in the available machine translation (MT) systems and their underlying approach based on word-by-word translation. An integrated functional approach to translating business texts is suggested on the basis of analyzing semantic and morphological features of actual text content and also on axiological and epistemic semantic features that bring to light subjective modality. The suggested technique is used to develop an algorithm of business text MT that makes it possible to resolve the word-by-word translation issue and conveys the meanings of short texts. Cases of testing the suggested technique and the derived algorithm are considered for the Russian–English language pair.