Multilingual sentiment analysis: from formal to informal and scarce resource languages

The ability to analyse online user-generated content related to sentiments (e.g., thoughts and opinions) on products or policies has become a de-facto skillset for many companies and organisations. Besides the challenge of understanding formal textual content, it is also necessary to take into consideration the informal and mixed linguistic nature of online social media languages, which are often coupled with localised slang as a way to express ‘true’ feelings. Due to the multilingual nature of social media data, analysis based on a single official language may carry the risk of not capturing the overall sentiment of online content. While efforts have been made to understand multilingual sentiment analysis based on a range of informal languages, no significant electronic resource has been built for these localised languages. This paper reviews the various current approaches and tools used for multilingual sentiment analysis, identifies challenges along this line of research, and provides several recommendations including a framework that is particularly applicable for dealing with scarce resource languages.     

A survey on sentiment analysis and opinion mining for social multimedia

Social media sentiment analysis (also known as opinion mining) which aims to extract people’s opinions, attitudes and emotions from social networks has become a research hotspot. Conventional sentiment analysis concentrates primarily on the textual content. However, multimedia sentiment analysis has begun to receive attention since visual content such as images and videos is becoming a new medium for self-expression in social networks. In order to provide a reference for the researchers in this active area, we give an overview of this topic and describe the algorithms of sentiment analysis and opinion mining for social multimedia. Having conducted a brief review on textual sentiment analysis for social media, we present a comprehensive survey of visual sentiment analysis on the basis of a thorough investigation of the existing literature. We further give a summary of existing studies on multimodal sentiment analysis which combines multiple media channels. We finally summarize the existing benchmark datasets in this area, and discuss the future research trends and potential directions for multimedia sentiment analysis. This survey covers 100 articles during 2008–2018 and categorizes existing studies according to the approaches they adopt.

     

A bilingual approach for conducting Chinese and English social media sentiment analysis

Due to the advancement of technology and globalization, it has become much easier for people around the world to express their opinions through social media platforms. Harvesting opinions through sentiment analysis from people with different backgrounds and from different cultures via social media platforms can help modern organizations, including corporations and governments understand customers, make decisions, and develop strategies. However, multiple languages posted on many social media platforms make it difficult to perform a sentiment analysis with acceptable levels of accuracy and consistency. In this paper, we propose a bilingual approach to conducting sentiment analysis on both Chinese and English social media to obtain more objective and consistent opinions. Instead of processing English and Chinese comments separately, our approach treats review comments as a stream of text containing both Chinese and English words. That stream of text is then segmented by our segment model and trimmed by the stop word lists which include both Chinese and English words. The stem words are then processed into feature vectors and then applied with two exchangeable natural language models, SVM and N-Gram. Finally, we perform a case study, applying our proposed approach to analyzing movie reviews obtained from social media. Our experiment shows that our proposed approach has a high level of accuracy and is more effective than the existing learning-based approaches.     

基于深度学习的社交网络平台细粒度情感分析

文本语言的情感分析历来是自然语言处理领域的热点研究课题,尤其是在当下互联网迈入web2.0时代,多样的社交网络平台呈现出巨量而丰富的文本情感信息,因此挖掘网络数据文本信息并作情感倾向判断对人机交互与人工智能具有重大的现实意义。传统的解决文本情感分析问题的方法主要是浅层学习算法,利用回归、分类等方案实现特征的提取及分类。以这类方法为起点,本文探索采用深度学习的方法对网络文本进行细粒度的情感分析,以期达到即时获取依附于网络世界的社会人的情感,甚至是让机器达到对人类情感表达的深度理解。对于深度学习的具体实现,本文采用的是降噪自编码器来对文本进行无标记特征学习并进行情感分类,后文中利用实验训练获得最佳的参数设置,并通过对实验结果的分析和评估论证深度学习对于情感信息的强大解析能力。     

A weakly-supervised extractive framework for sentiment-preserving document summarization

The popularity of social media sites provides new ways for people to share their experiences and convey their opinions, leading to an explosive growth of user-generated content. Text data, owing to the amazing expressiveness of natural language, is of great value for people to explore various kinds of knowledge. However, considerable user-generated text contents are longer than what a reader expects, making automatic document summarization a necessity to facilitate knowledge digestion. In this paper, we focus on the reviews-like sentiment-oriented textual data. We propose the concept of Sentiment-preserving Document Summarization (SDS), aiming at summarizing a long textual document to a shorter version while preserving its main sentiments and not sacrificing readability. To tackle this problem, using deep neural network-based models, we devise an end-to-end weakly-supervised extractive framework, consisting of a hierarchical document encoder, a sentence extractor, a sentiment classifier, and a discriminator to distinguish the extracted summaries from the natural short reviews. The framework is weakly-supervised in that no ground-truth summaries are used for training, while the sentiment labels are available to supervise the generated summary to preserve the sentiments of the original document. In particular, the sentence extractor is trained to generate summaries i) making the sentiment classifier predict the same sentiment category as the original longer documents, and ii) fooling the discriminator into recognizing them as human-written short reviews. Experimental results on two public datasets validate the effectiveness of our framework.

     

情感分析与认知

分析了情感分析的3个主要步骤,包括文本情感获取与表达、文本情感分类与计算以及文本情感分析的应用.情感分析得到的结论主要是对相关观点的摘要、对相关事件态度的预测或者统计等,但这些结论都没有发挥文本情感在认知中的作用.为了将情感分析应用于认知科学,提出了情感由情感信号和情感实体组成的观点.情感信号主要是指情感的一些形式载体,比如心跳加速、脸红等这些人体内外的某些表现,表达情感的文字、图片、声音等这类媒体.情感实体主要是指人类对情感形成的一种共识,比如爱、恨、憎恶、高兴、羞愧、嫉妒、内疚、恐惧、焦虑等与人的意识相关联的部分.同时提出了在人工智能中利用情感信息的设想.这对于模拟情感对认知的影响具有一定的意义.     

Regional Sentiment Bias in Social Media Reporting During Crises

Crisis events such as terrorist attacks are extensively commented upon on social media platforms such as Twitter. For this reason, social media content posted during emergency events is increasingly being used by news media and in social studies to characterize the public’s reaction to those events. This is typically achieved by having journalists select ‘representative’ tweets to show, or a classifier trained on prior human-annotated tweets is used to provide a sentiment/emotion breakdown for the event. However, social media users, journalists and annotators do not exist in isolation, they each have their own context and world view. In this paper, we ask the question, ‘to what extent do local and international biases affect the sentiments expressed on social media and the way that social media content is interpreted by annotators’. In particular, we perform a multi-lingual study spanning two events and three languages. We show that there are marked disparities between the emotions expressed by users in different languages for an event. For instance, during the 2016 Paris attack, there was 16% more negative comments written in the English than written in French, even though the event originated on French soil. Furthermore, we observed that sentiment biases also affect annotators from those regions, which can negatively impact the accuracy of social media labelling efforts. This highlights the need to consider the sentiment biases of users in different countries, both when analysing events through the lens of social media, but also when using social media as a data source, and for training automatic classification models.     

Discovering public sentiment in social media for predicting stock movement of publicly listed companies

The popularity of many social media sites has prompted both academic and practical research on the possibility of mining social media data for the analysis of public sentiment. Studies have suggested that public emotions shown through Twitter could be well correlated with the Dow Jones Industrial Average. However, it remains unclear how public sentiment, as reflected on social media, can be used to predict stock price movement of a particular publicly-listed company. In this study, we attempt to fill this research void by proposing a technique, called SMeDA-SA, to mine Twitter data for sentiment analysis and then predict the stock movement of specific listed companies. For the purpose of experimentation, we collected 200 million tweets that mentioned one or more of 30 companies that were listed in NASDAQ or the New York Stock Exchange. SMeDA-SA performs its task by first extracting ambiguous textual messages from these tweets to create a list of words that reflects public sentiment. SMeDA-SA then made use of a data mining algorithm to expand the word list by adding emotional phrases so as to better classify sentiments in the tweets. With SMeDA-SA, we discover that the stock movement of many companies can be predicted rather accurately with an average accuracy over 70%. This paper describes how SMeDA-SA can be used to mine social media date for sentiments. It also presents the key implications of our study.     

Effective deep learning based multimodal sentiment analysis from unstructured big data

More recently, as images, memes and graphics interchange formats have dominated social feeds, typographic/infographic visual content has emerged as an important social media component. This multimodal text combines text and image, defining a novel visual language that must be analysed because it has the potential to modify, confirm or grade the sentiment's polarity. The problem is how to effectively use information from the visual and textual content in image-text posts. This article presents a new deep learning-based multimodal sentiment analysis (MSA) model using multimodal data such as images, text and multimodal text (image with embedded text). The text analytic unit, the discretization control unit, the picture analytic component and the decision-making component are all included in this system. The discretization unit separates the text from the picture using the variant and channel augmented maximally stable extremal regions (VCA-MSERs) technique, which are then analysed as discrete elements and fed into the appropriate image and text analytics units. The text analytics system utilizes a stacked recurrent neural network with multilevel attention and feedback module (SRNN-MAFM) to detect the sentiment of the text. A deep convolutional neural network (CNN) structure with parallel-dilated convolution and self-attention module (PDC-SAM) is developed to forecast the emotional response to visual content. Finally, the decision component employs a Boolean framework including an OR function to evaluate and classify the output into three fine-grained sentiment classes: positive, neutral and negative. The proposed work is simulated in the python platform using the STS-Gold, Flickr 8k and B-T4SA datasets for sentiment analysis of text and visual and multimodal text. Simulation outcomes proved that the suggested method achieved better accuracy of 97.8%, 97.7% and 90% for text, visual and MSA individually compared to other methods.     

Unsupervised method for sentiment analysis in online texts

In recent years, the explosive growth of online media, such as blogs and social networking sites, has enabled individuals and organizations to write about their personal experiences and express opinions. Classifying these documents using a polarity metric is an arduous task. We propose a novel approach to predicting sentiment in online textual messages such as tweets and reviews, based on an unsupervised dependency parsing-based text classification method that leverages a variety of natural language processing techniques and sentiment features primarily derived from sentiment lexicons. These lexicons were created by means of a semiautomatic polarity expansion algorithm in order to improve accuracy in specific application domains. The results obtained for the Cornell Movie Review, Obama-McCain Debate and SemEval-2015 datasets confirm the competitive performance and the robustness of the system.     

Phonetic-enriched text representation for Chinese sentiment analysis with reinforcement learning

The Chinese pronunciation system offers two characteristics that distinguish it from other languages: deep phonemic orthography and intonation variations. In this paper, we hypothesize that these two important properties can play a major role in Chinese sentiment analysis. In particular, we propose two effective features to encode phonetic information and, hence, fuse it with textual information. With this hypothesis, we propose Disambiguate Intonation for Sentiment Analysis (DISA), a network that we develop based on the principles of reinforcement learning. DISA disambiguates intonations for each Chinese character (pinyin) and, hence, learns precise phonetic representations. We also fuse phonetic features with textual and visual features to further improve performance. Experimental results on five different Chinese sentiment analysis datasets show that the inclusion of phonetic features significantly and consistently improves the performance of textual and visual representations and surpasses the state-of-the-art Chinese character-level representations.     

面向产品评论分析的短文本情感主题模型

情感主题联合生成模型已经成功应用于网络评论分析.然而,随着智能终端设备的广泛应用,由于屏幕及输入限制,用户书写的评论越来越短,我们不得不面对短评论中的文本稀疏问题.本文提出了一个针对短文本的联合情感--主题模型SSTM（Short-text sentiment-topic model）来解决稀疏性问题.不同于一般主题模型中通常采用的基于文档产生过程的建模方法,我们直接对整个语料集合的产生过程建模.在产生文档集的过程中,我们每次采样一个词对,同一个词对中的词有相同的情感极性和主题.我们将SSTM模型应用于两个真实网络评论数据集.在三个实验任务中,通过定性分析验证了主题发现的有效性,并与经典方法进行定量对比,SSTM模型的文档级情感分类性能也有较大提升.     

Sentiment analysis leveraging emotions and word embeddings

Sentiment analysis and opinion mining are valuable for extraction of useful subjective information out of text documents. These tasks have become of great importance, especially for business and marketing professionals, since online posted products and services reviews impact markets and consumers shifts. This work is motivated by the fact that automating retrieval and detection of sentiments expressed for certain products and services embeds complex processes and pose research challenges, due to the textual phenomena and the language specific expression variations. This paper proposes a fast, flexible, generic methodology for sentiment detection out of textual snippets which express people’s opinions in different languages. The proposed methodology adopts a machine learning approach with which textual documents are represented by vectors and are used for training a polarity classification model. Several documents’ vector representation approaches have been studied, including lexicon-based, word embedding-based and hybrid vectorizations. The competence of these feature representations for the sentiment classification task is assessed through experiments on four datasets containing online user reviews in both Greek and English languages, in order to represent high and weak inflection language groups. The proposed methodology requires minimal computational resources, thus, it might have impact in real world scenarios where limited resources is the case.     

基于卷积神经网络的图文融合媒体情感预测

近年来,用户在社交媒体上越来越多地使用多媒体内容来分享经历和表达情绪。相比单独的文本和图像,融合文本和图像的多媒体内容能够更为充分地揭示用户的真实情感。针对单一文本或图像的情感不明显问题,提出了一种基于卷积神经网络(CNN)的图文融合媒体的情感分析方法。该方法融合图像特征与三个不同级别(词语级、短语级和句子级)的文本特征构建CNN模型,以分析比较不同层次的语义特征对情感预测的影响。在真实数据集上的实验结果表明,通过捕捉文本情感特征和图像情感特征之间的内部联系,可以更准确地实现对图文融合媒体情感的预测。     

Joint sentiment/topic modeling on text data using a boosted restricted Boltzmann Machine

Recently by the development of the Internet and the Web, different types of social media such as web blogs become an immense source of text data. Through the processing of these data, it is possible to discover practical information about different topics, individual’s opinions and a thorough understanding of the society. Therefore, applying models which can automatically extract the subjective information from documents would be efficient and helpful. Topic modeling methods and sentiment analysis are the raised topics in natural language processing and text mining fields. In this paper a new structure for joint sentiment-topic modeling based on a Restricted Boltzmann Machine (RBM) which is a type of neural networks is proposed. By modifying the structure of RBM as well as appending a layer which is analogous to sentiment of text data to it, we propose a generative structure for joint sentiment topic modeling based on neural networks. The proposed method is supervised and trained by the Contrastive Divergence algorithm. The new attached layer in the proposed model is a layer with the multinomial probability distribution which can be used in text data sentiment classification or any other supervised application. The proposed model is compared with existing models in the experiments such as evaluating as a generative model, sentiment classification, information retrieval and the corresponding results demonstrate the efficiency of the method.

     

短文本情感分析的研究现状 ——从社交媒体到资源稀缺语言

随着社交网络的逐渐成熟,各类语种的文本出现在社交网络上。而这些非规范的短文本蕴藏着人们对事物的褒贬、需求等意见,是国家政府和企业了解公众舆论的重要参考信息,具有重大的研究价值和应用价值。首先,对 目前互联网短文本情感分析领域常用的神经网络、跨语言和应用语言学知识等研究方法进行归纳和总结;其次,对当前短文本情感分析研究的热点领域——社交媒体和资源稀缺语言的情感分析进行现状分析;最后,对短文本情感分析研究的趋势进行总结,分析存在的问题,并对未来进行展望。     

Sentimental causal rule discovery from Twitter

Social media, especially Twitter is now one of the most popular platforms where people can freely express their opinion. However, it is difficult to extract important summary information from many millions of tweets sent every hour. In this work we propose a new concept, sentimental causal rules, and techniques for extracting sentimental causal rules from textual data sources such as Twitter which combine sentiment analysis and causal rule discovery. Sentiment analysis refers to the task of extracting public sentiment from textual data. The value in sentiment analysis lies in its ability to reflect popularly voiced perceptions that are stated in natural language. Causal rules on the other hand indicate associations between different concepts in a context where one (or several concepts) cause(s) the other(s). We believe that sentimental causal rules are an effective summarization mechanism that combine causal relations among different aspects extracted from textual data as well as the sentiment embedded in these causal relationships. In order to show the effectiveness of sentimental causal rules, we have conducted experiments on Twitter data collected on the Kurdish political issue in Turkey which has been an ongoing heated public debate for many years. Our experiments on Twitter data show that sentimental causal rule discovery is an effective method to summarize information about important aspects of an issue in Twitter which may further be used by politicians for better policy making.     

Survey of visual sentiment prediction for social media analysis

Recent years have witnessed a rapid spread of multi-modality microblogs like Twitter and Sina Weibo composed of image, text and emoticon. Visual sentiment prediction of such microblog based social media has recently attracted ever-increasing research focus with broad application prospect. In this paper, we give a systematic review of the recent advances and cutting-edge techniques for visual sentiment analysis. To this end, in this paper we review the most recent works in this topic, in which detailed comparison as well as experimental evaluation are given over the cutting-edge methods. We further reveal and discuss the future trends and potential directions for visual sentiment prediction.     

Actionable Social Media Competitive Analytics For Understanding Customer Experiences

A large amount of user-generated content is now freely available on social media sites. To increase their competitive advantage, companies need to monitor and analyze not only the customer-generated content on their own social media sites, but also the content on their competitors’ social media sites. In this article, we describe a framework to integrate several techniques including quantitative analysis, text mining, and sentiment analysis for analyzing and comparing social media content from business competitors. Specifically, we conducted an in-depth case study which applies our developed framework to the analysis and comparison of social media content on the Facebook sites of the three largest drugstore chains in the United States: Walgreens, CVS, and Rite Aid. We found similarities and differences in the social media use among the three drugstore chains. We discuss the implications of our findings and provide recommendations to help companies develop their social media competitive analysis strategies.