DT-FNN based effective hybrid classification scheme for twitter sentiment analysis

Multimedia Tools and Applications - Sentiment analysis refers to the interpretation and computational study of emotions, opinions and appraisals within the text data using text analysis methods. A...     

基于语义特征的多视图情感分类方法

情感分析也称为意见挖掘,是对文本中所包含的情感倾向进行分析的技术。目前很多情感分析工作都是基于纯文本的。而在微博上,除了文本,大量的图片信息也蕴含了丰富的情感信息。本文提出了一种基于文本和图像的多模态分类算法,通过使用潜在语义分析,将文本特征和图像特征分别映射到同维度下的语义空间,得到各自的语义特征,并用SVM-2K进行分类。利用新浪微博热门微博栏目下爬取的文字和配图的微博数据进行了实验。实验结果表明,通过融合文本和图像的语义特征,情感分类的效果好于单独使用文本特征或者图像特征。     

一种融合上下文特征的中文隐式情感分类模型

对网络上海量的文本数据进行情感分析,可以更好地挖掘网民行为规律、帮助决策机构了解舆情倾向和改善商家服务质量。在实际表达中,人们除了采用带有明显情感词的主观表达外,还采用含蓄的方式表达自己的主观倾向。带有显式情感词的文本情感分析作为自然语言处理领域的基础性研究任务,已经取得了丰富的研究成果。然而,针对隐式文本的情感分析技术还处于起步阶段。与显式情感分析任务相比,隐式情感分类任务更加困难。隐式表达文本具有中立性表达、缺乏情感词和上下文依赖的特点,使得传统的文本分类方法不再适用。针对以上问题,采用word2vec词嵌入技术提取文本特征,分别进行了基于TextCNN、LSTM和BiGRU分类模型的研究。在各个深度分类模型研究基础上,还进行了融合注意力机制的分类模型研究。针对隐式表达对上下文内容依赖的特点,设计了一种融合上下文语义特征和注意力机制的分类模型,增强了部分中立性隐式表达句的分类效果。最后在SMP2019公开数据集上进行了实验,取得了比上述几种基础深度网络模型与融合注意力机制分类模型更好的分类效果。     

基于本体特征的影评细粒度情感分类

针对中文影评情感分类中缺少特征属性及情感强度层面的粒度划分问题,提出一种基于本体特征的细粒度情感分类模型。首先,利用词频逆文档频率(TF-IDF)和TextRank算法提取电影特征,构建本体概念模型。其次,将电影特征属性和普鲁契克多维度情绪模型与双向长短时记忆网络(Bi-LSTM)融合,构建了在特征粒度层面和八分类情感强度下的细粒度情感分类模型。实验中,本体特征分析表明:观影人对故事属性关注度最高,继而是题材、人物、场景、导演等特征;模型性能分析表明:基于特征粒度和八分类情感强度,与应用情感词典、机器学习、Bi-LSTM网络算法在整体粒度和三分类情感强度层面的其他5个分类模型相比,该模型不仅有较高的F1值(0. 93),而且还能提供观影人对电影属性的情感偏好和情感强度参考,实现了中文影评更细粒度的情感分类。     

基于特征情感色彩的文档流派分类研究

流派分类和基于主题的文本分类最大的区别之处就在于文本的特征。流派分类需要能够描述文档风格的、表达更强语义信息的特征,基于特征情感色彩的分类方法是将情感色彩这种语义信息附加到特征上。首先介绍了文档流派分类的概念及其应用,然后分析了流派分类的文本特征和词汇的情感倾向权值的几种计算方法,论述了基于特征情感色彩的文档流派分类过程,最后对几种分类方法进行了实验结果分析和比较。     

KL-NF technique for sentiment classification

This work proposes sentiment analysis for low-resource languages like Hindi using Neuro-Fuzzy Technique. Low-resource languages suffer from the scarcity of resources; consequently, we propose a method that can be implemented for any language. We use information theory for establishing a relation between terms that exists in a sentence. This work proposes a novel approach for calculating feature values using Kullback-Leibler (KL) divergence method. The feature values are employed to calculate the membership values associated with the Fuzzy logic in Neuro-Fuzzy Technique. The novelty of this method lies in its predictive nature that can mitigate the impact generated from un-labeled, unknown data or multi-domain data. We have seen the results for multi-domain data in our experiments. We evaluate our results using Accuracy, Precision, Recall and F1-Score. Our experiments show the efficacy of the proposed approach. It achieved 93.01% accuracy for English dataset and 91.18% accuracy for Hindi dataset which is more than the other state-of-art techniques like Naïve Bayes and SVM. Additionally, we found that our approach provides satisfactory results with multi-domain data as both the datasets were of different domains.

     

Exploiting semantic relationships for unsupervised expansion of sentiment lexicons

The literature in sentiment analysis has widely assumed that semantic relationships between words cannot be effectively exploited to produce satisfactory sentiment lexicon expansions. This assumption stems from the fact that words considered to be “close” in a semantic space (e.g., word embeddings) may present completely opposite polarities, which might suggest that sentiment information in such spaces is either too faint, or at least not readily exploitable. Our main contribution in this paper is a rigorous and robust challenge to this assumption: by proposing a set of theoretical hypotheses and corroborating them with strong experimental evidence, we demonstrate that semantic relationships can be effectively used for good lexicon expansion. Based on these results, our second contribution is a novel, simple, and yet effective lexicon-expansion strategy based on semantic relationships extracted from word embeddings. This strategy is able to substantially enhance the lexicons, whilst overcoming the major problem of lexicon coverage. We present an extensive experimental evaluation of sentence-level sentiment analysis, comparing our approach to sixteen state-of-the-art (SOTA) lexicon-based and five lexicon expansion methods, over twenty datasets. Results show that in the vast majority of cases our approach outperforms the alternatives, achieving coverage of almost 100% and gains of about 26% against the best baselines. Moreover, our unsupervised approach performed competitively against SOTA supervised sentiment analysis methods, mainly in scenarios with scarce information. Finally, in a cross-dataset comparison, our approach turned out to be as competitive as (i.e., statistically tie with) state-of-the-art supervised solutions such as pre-trained transformers (BERT), even without relying on any training (labeled) data. Indeed in small datasets or in datasets with scarce information (short messages), our solution outperformed the supervised ones by large margins.     

基于时序和上下文特征的中文隐式情感分类模型

对社交网络上的海量文本信息进行情感分析可以更好地挖掘网民行为规律,从而帮助决策机构了解舆情倾向以及帮助商家改善服务质量。由于不存在关键情感特征、表达载体形式和文化习俗等因素的影响,中文隐式情感分类任务比其他语言更加困难。已有的中文隐式情感分类方法以卷积神经网络（CNN）为主,这些方法存在着无法获取词语的时序信息和在隐式情感判别中未合理利用上下文情感特征的缺陷。为了解决以上问题,采用门控卷积神经网络（GCNN）提取隐式情感句的局部重要信息,采用门控循环单元（GRU）网络增强特征的时序信息;而在隐式情感句的上下文特征处理上,采用双向门控循环单元（BiGRU）+注意力机制（Attention）的组合提取重要情感特征;在获得两种特征后,通过融合层将上下文重要特征融入到隐式情感判别中;最后得到的融合时序和上下文特征的中文隐式情感分类模型被命名为GGBA。在隐式情感分析评测数据集上进行实验,结果表明所提出的GGBA模型在宏平均准确率上比普通的文本CNN即TextCNN提高了3.72%、比GRU提高了2.57%、比中断循环神经网络（DRNN）提高了1.90%,由此可见, GGBA模型在隐式情感分析任务中比基础模型获得了更好的分类性能。     

An empirical study of sentiment analysis for chinese documents

Up to now, there are very few researches conducted on sentiment classification for Chinese documents. In order to remedy this deficiency, this paper presents an empirical study of sentiment categorization on Chinese documents. Four feature selection methods (MI, IG, CHI and DF) and five learning methods (centroid classifier, K-nearest neighbor, winnow classifier, Naïve Bayes and SVM) are investigated on a Chinese sentiment corpus with a size of 1021 documents. The experimental results indicate that IG performs the best for sentimental terms selection and SVM exhibits the best performance for sentiment classification. Furthermore, we found that sentiment classifiers are severely dependent on domains or topics.     

Hybrid sentiment classification on twitter aspect-based sentiment analysis

Social media sites and applications, including Facebook, YouTube, Twitter and blogs, have become major social media attractions today. The huge amount of information from this medium has become an attractive resource for organisations to monitor the opinions of users, and therefore, it is receiving a lot of attention in the field of sentiment analysis. Early work on sentiment analysis approached this problem at a document-level, where the overall sentiment was identified, rather than the details of the sentiment. This research took into account the use of an aspect-based sentiment analysis on Twitter in order to perform a finer-grained analysis. A new hybrid sentiment classification for Twitter is proposed by embedding a feature selection method. A comparison of the accuracy of the classification by the principal component analysis (PCA), latent semantic analysis (LSA), and random projection (RP) feature selection methods are presented in this paper. Furthermore, the hybrid sentiment classification was validated using Twitter datasets to represent different domains, and the evaluation with different classification algorithms also demonstrated that the new hybrid approach produced meaningful results. The implementations showed that the new hybrid sentiment classification was able to improve the accuracy performance from the existing baseline sentiment classification methods by 76.55, 71.62 and 74.24%, respectively.     

A valences-totaling model for English sentiment classification

Sentiment classification plays an important role in everyday life, in political activities, activities of commodity production and commercial activities. Finding a time-effective and highly accurate solution to the classification of emotions is challenging. Today, there are many models (or methods) to classify the sentiment of documents. Sentiment classification has been studied for many years and is used widely in many different fields. We propose a new model, which is called the valences-totaling model (VTM), by using cosine measure (CM) to classify the sentiment of English documents. VTM is a new model for English sentiment classification. In this study, CM is a measure of similarity between two words and is used to calculate the valence (and polarity) of English semantic lexicons. We prove that CM is able to identify the sentiment valence and the sentiment polarity of the English sentiment lexicons online in combination with the Google search engine with AND operator and OR operator. VTM uses many English semantic lexicons. These English sentiment lexicons are calculated online and are based on the Internet. We present a full range of English sentences; thus, the emotion expressed in the English text is classified with more precision. Our new model is not dependent on a special domain and training data set—it is a domain-independent classifier. We test our new model on the Internet data in English. The calculated valence (and polarity) of English semantic words in this model is based on many documents on millions of English Web sites and English social networks.     

基于上下文感知的方面类别情感分类

由于一个评论往往会涉及多种方面类别及情感倾向,而传统注意力机制难以区分方面词和情感词的对应关系,从而影响评论同时存在多种方面类别时的情感极性分析.为了解决上述问题,提出了一种基于上下文感知的方面类别情感分类模型(MA-DSA).该模型通过重构方面向量捕获句子中更多样且有效的语义特征,并将其融入上下文向量,然后将上下文向量通过DiSA模块进一步捕捉句子内部情感特征,确定方面词与情感词的关系,进而对指定方面类别进行情感分类.在SemEval的三个数据集上的实验结果表明,MA-DSA模型在Restaurant-2014数据集上的三个指标值均优于基准模型,证明了该模型的有效性.     

基于句法增强的细粒度情感三元组抽取方法

属性级情感三元组抽取（aspect sentiment triplet extraction,ASTE）任务主要是从句子中检测出属性词及其对应的评价词和情感倾向,然而当抽取多词属性词和评价词时,无法准确地抽取出全部的单词;当面对重复的属性词和评价词时,以往的研究很难学习到\     

层次化双注意力神经网络模型的情感分析研究

在篇章级的情感分类中由于篇章级文本较长,特征提取较普通句子级分析相对较难,大多方法使用层次化的模型进行篇章文本的情感分析,但目前的层次化模型多以循环神经网络和注意力机制为主,单一的循环神经网络结构提取的特征不够明显。本文针对篇章级的情感分类任务,提出一种层次化双注意力神经网络模型。首先对卷积神经网络进行改进,构建词注意力卷积神经网络。然后模型从两个层次依次提取篇章特征,第一层次使注意力卷积神经网络发现每个句子中的重要词汇,提取句子的词特征,构建句子特征向量;第二层次以循环神经网络获取整个篇章的语义表示,全局注意力机制发现篇章中每个句子的重要性,分配以不同的权重,最后构建篇章的整体语义表示。在IMDB、YELP 2013、YELP 2014数据集上的实验表明,模型较当前最好的模型更具优越性。     

Ensemble of feature sets and classification algorithms for sentiment classification

In this paper, we make a comparative study of the effectiveness of ensemble technique for sentiment classification. The ensemble framework is applied to sentiment classification tasks, with the aim of efficiently integrating different feature sets and classification algorithms to synthesize a more accurate classification procedure. First, two types of feature sets are designed for sentiment classification, namely the part-of-speech based feature sets and the word-relation based feature sets. Second, three well-known text classification algorithms, namely na?¨ve Bayes, maximum entropy and support vector machines, are employed as base-classifiers for each of the feature sets. Third, three types of ensemble methods, namely the fixed combination, weighted combination and meta-classifier combination, are evaluated for three ensemble strategies. A wide range of comparative experiments are conducted on five widely-used datasets in sentiment classification. Finally, some in-depth discussion is presented and conclusions are drawn about the effectiveness of ensemble technique for sentiment classification.     

A two-stage framework for cross-domain sentiment classification

Supervised sentiment classification systems are typically domain-specific, and the performance decreases sharply when transferred from one domain to another domain. Building these systems involves annotating a large amount of data for every domain, which needs much human labor. So, a reasonable way is to utilize labeled data in one existed (or called source) domain for sentiment classification in target domain. To address this problem, we propose a two-stage framework for cross-domain sentiment classification. At the “building a bridge” stage, we build a bridge between the source domain and the target domain to get some most confidently labeled documents in the target domain; at the “following the structure” stage, we exploit the intrinsic structure, revealed by these most confidently labeled documents, to label the target-domain data. The experimental results indicate that the proposed approach could improve the performance of cross-domain sentiment classification dramatically.     

中文微博情感分类的简单多标签排序算法

针对中文微博文本情感分类中每个样本最多只有两种有序情感标签的情形,提出了一种简单的多标签排序算法——TSMLR,该算法采用两步学习和两步分类的策略,通过学习情感标签之间的主次关系,对微博文本的情感进行分类并对情感标签进行排序。首先,将一个多标签排序问题转化为八个多类单标签分类问题,分别对主要情感标签和次要情感标签进行学习;然后,利用得到的分类模型对微博表达的情感进行两步分类,首先给出主要情感标签,再给出次要情感标签。通过在NLP&CC2014的中文微博文本情感分析评测数据集上进行实验,与校准标签排序方法(CLR)相比,TSMLR方法的准确度和平均精度分别提高了8.59%和9.28%,1-错误率相应下降了9.77%,而且TSMLR所需的训练时间相对较少。实验结果表明:TSMLR对标签之间顺序关系的学习能够有效提高对中文微博情感分类的准确率。     

基于情感时序距离和转折同化的文本情感分类

考虑到中文评价文本的整体情感倾向性与其表达的情感顺序有很大关系,且在具有情感倾向的中文文本中,越是靠近文本最后所表达的情感倾向,对于整个文本的情感分类影响越大。因此对于情感倾向表达不明显或者表达不单一的短文本,通过考虑文本中情感节点出现的顺序以及情感转折同化来对文本进行情感分类。在来自某购物网站爬取的中评评价文本数据集上的实验结果显示,提出的分类方法明显高于单纯基于词特征的支持向量机（SVM）分类器。     

基于汉语情感词表的句子情感倾向分类研究

提出了一种基于汉语情感词词表的加权线性组合的句子情感分类方法。该方法通过已有的五种资源构建了中文情感词词表,并采用加权线性组合的句子情感分类方法对句子进行情感类别判断。实验结果表明,直接利用词汇语言粒度的句子情感分类综合F值为78.62％,若加入了否定短语语言粒度后,句子情感分类的综合F值提高了4.14％。