To be rational or sensitive? The gender difference in how textual environment cue and personal characteristics influence the sentiment expression on social media

Online sentiments expressed by users play critical roles in various social media-based applications, and thus understanding the mechanism of what determines users expressing sentiment with different polarities bears strategic importance. Based on the affective response model (ARM), we develop a conceptual model about the determinants of users’ online sentiment polarity, from the cues of the textual environment from the target tweet and the user’s personal characteristics. Furthermore, the role of gender difference in these effects is also included. Empirical results indicated that users with higher social interactivity and positive historical sentiment expression are more likely to express positive sentiment towards online tweets with higher positive sentiment intensity. Females are more sensitive to the cue of textual environment, i.e, sentiment intensity, in the target tweets when expressing sentiments, while males are more rational when expressing online sentiment than females. Our study supplements the existing study on users’ online interaction behavior as rational and affective action by introducing a new way to study the driving behavior of sentiment expression.     

FEROM: Feature Extraction and Refinement for Opinion Mining

Opinion mining involves the analysis of customer opinions using product reviews and provides meaningful information including the polarity of the opinions. In opinion mining, feature extraction is important since the customers do not normally express their product opinions holistically but separately according to its individual features. However, previous research on feature‐based opinion mining has not had good results due to drawbacks, such as selecting a feature considering only syntactical grammar information or treating features with similar meanings as different. To solve these problems, this paper proposes an enhanced feature extraction and refinement method called FEROM that effectively extracts correct features from review data by exploiting both grammatical properties and semantic characteristics of feature words and refines the features by recognizing and merging similar ones. A series of experiments performed on actual online review data demonstrated that FEROM is highly effective at extracting and refining features for analyzing customer review data and eventually contributes to accurate and functional opinion mining.     

Extracting Multiword Sentiment Expressions by Using a Domain‐Specific Corpus and a Seed Lexicon

This paper presents a novel approach to automatically generate Korean multiword sentiment expressions by using a seed sentiment lexicon and a large‐scale domain‐specific corpus. A multiword sentiment expression consists of a seed sentiment word and its contextual words occurring adjacent to the seed word. The multiword sentiment expressions that are the focus of our study have a different polarity from that of the seed sentiment word. The automatically extracted multiword sentiment expressions show that 1) the contextual words should be defined as a part of a multiword sentiment expression in addition to their corresponding seed sentiment word, 2) the identified multiword sentiment expressions contain various indicators for polarity shift that have rarely been recognized before, and 3) the newly recognized shifters contribute to assigning a more accurate polarity value. The empirical result shows that the proposed approach achieves improved performance of the sentiment analysis system that uses an automatically generated lexicon.     

面向顾客点评数据的属性层次观点挖掘研究

This paper focuses on how to improve aspect-level opinion mining for online customer reviews. We first propose a novel generative topic model, the Joint Aspect/Sen-timent (JAS) model, to jointly extract aspects and aspect-dependent sentiment lexicons from online customer reviews. An as-pect-dependent sentiment lexicon refers to the aspect-specific opinion words along with their aspect-aware sentiment polarities with respect to a specific aspect. We then apply the extracted aspect- dependent sentiment lexi-cons to a series of aspect-level opinion mining tasks, including implicit aspect identification, aspect-based extractive opinion summarization, and aspect-level sentiment classification. Experimental results demonstrate the effectiveness of the JAS model in learning aspect- dependent sentiment lexicons and the practical values of the extracted lexicons when applied to these practical tasks.     

Motives,frequency and attitudes toward emoji and emoticon use

Electronic Mediated Communication (EMC) has become highly prevalent in our daily lives. Many of the communication formats used in EMC are text-based (e.g., instant messaging), and users often include visual paralinguistic cues in their messages. In the current study, we examined the usage of two such cues – emoji and emoticons. Specifically, we compared self-reported frequency of use, as well as attitudes (6 bipolar items, e.g., “fun” vs. “boring”) and motives for their usage (9 motives, e.g., “express how I feel to others”). We also examined these indicators according to age and gender. Overall, participants (N?=?474, 72.6% women; M_age?=?30.71, SD?=?12.58) reported using emoji (vs. emoticons) more often, revealed more positive attitudes toward emoji usage, and identified more with motives to use them. Moreover, all the ratings were higher among younger (vs. older) participants. Results also showed that women reported to use emoji (but not emoticons) more often and expressed more positive attitudes toward their usage than men. However, these gender differences were particularly evident for younger participants. No gender differences were found for emoticons usage. These findings add to the emerging body of literature by showing the relevance of considering age and gender, and their interplay, when examining patterns of emoji and emoticons use.     

Sarcastic sentiment detection in tweets streamed in real time: a big data approach

Sarcasm is a type of sentiment where people express their negative feelings using positive or intensified positive words in the text. While speaking, people often use heavy tonal stress and certain gestural clues like rolling of the eyes, hand movement, etc. to reveal sarcastic. In the textual data, these tonal and gestural clues are missing, making sarcasm detection very difficult for an average human. Due to these challenges, researchers show interest in sarcasm detection of social media text, especially in tweets. Rapid growth of tweets in volume and its analysis pose major challenges. In this paper, we proposed a Hadoop based framework that captures real time tweets and processes it with a set of algorithms which identifies sarcastic sentiment effectively. We observe that the elapse time for analyzing and processing under Hadoop based framework significantly outperforms the conventional methods and is more suited for real time streaming tweets.     

Uncovering sentiment and retweet patterns of disaster-related tweets from a spatiotemporal perspective – A case study of Hurricane Harvey

Social media has been widely used for emergency communication both in disaster-affected areas and unaffected areas. Comparing emotional reaction and information propagation between on-site users and off-site users from a spatiotemporal perspective can help better comprehend collective human behavior during natural disasters. In this study, we investigate sentiment and retweet patterns of disaster-affected areas and disaster-unaffected areas at different stages of Hurricane Harvey. The results show that off-site tweets were more negative than on-site tweets, especially during the disaster. As for retweet patterns, indifferent-neutral and positive tweets spread broader than mixed-neutral and negative tweets. However, negative tweets spread faster than positive tweets, which reveals that social media users were more sensitive to negative information in disaster situations. With the development of the disaster, social media users were more sensitive to on-site positive messages than off-site negative posts. This data-driven study reveals the significant effect of sentiment expression on the publication and re-distribution of disaster-related messages. It generates implications for emergency communication and disaster management.     

基于多维特征分析的社交网络意见领袖挖掘

在社交网络中进行意见领袖的挖掘对信息传播与演化的深度分析、舆情监控和引导具有重要意义,本文综合结构特征、行为特征和用户的情感特征对意见领袖节点挖掘问题进行研究.本文首先对微博真实文本数据进行话题识别得到主题社区,在主题社区中基于用户节点之间的关注关系构建交互网络拓扑.然后分别从结构、行为和情感三个维度对用户的影响力进行度量.最后,分析用户在主题社区中的影响力分布与传播规律,提出意见领袖识别算法MFP(Multi-Feature PageRank).实验表明,该算法可有效地挖掘潜在的意见领袖节点,能够获得较高的支持率.     

大数据环境下的网络舆情管理系统设计

随着网络的快速发展,越来越多的人们通过网络发表个人观点及看法,网络舆情成为社会舆情中的重点对象和主要方式.本文通过对大数据环境下网络舆情及其特点的阐述、分析,结合数据挖掘、文本情感分析等技术,初步构建出了网络舆情管理系统的模型.     

Detecting Streaming of Twitter Spam Using Hybrid Method

Twitter, the social network which evolving faster and regular usage by millions of people and who become addicted to it. So spam playing a major role for Twitter users to distract them and grab their attention over them. Spammers actually detailed like who send unwanted and irrelevant messages or websites and promote them to several users. To overcome the problem many researchers proposed some ideas using some machine learning algorithms to detect the spammers. In this research work, a new hybrid approach is proposed to detect the streaming of Twitter spam in a real-time using the combination of a Decision tree, Particle Swarm Optimization and Genetic algorithm. Twitter has given access to the researchers to get tweets from its Twitter-API for real-time streaming of tweet data which they can get direct access to public tweets. Here 600 million tweets are created by using URL based security tool and further some features are extracted for representation of tweets in real-time detection of spam. In addition, our research results are compared with other hybrid algorithms which a better detection rate is given by our proposed work.     

融入音乐子人格特质和社交网络行为分析的音乐推荐方法

为了可以实时推荐符合人们情感状态的音乐,提出了一种融入音乐子人格特质的社交网络行为分析的音乐推荐算法,该算法通过分析用户发表在微博等社交媒体上的状态,计算用户在该情感状态下对音乐的偏好程度;选择在该情感状态下音乐偏好相似的最近邻用户,最后融入音乐子人格特质进行偏好度计算,为用户推荐最适合其情感状态的音乐。实验结果表明,该算法可以缓解用户数据稀疏性对推荐结果的影响,能够提高推荐系统的推荐质量。     

多交互图卷积网络用于方面情感分析

方面情感分析旨在识别句子中特定方面的情感极性,是一项细粒度情感分析任务。传统基于注意力机制方法,仅在单词之间进行单一的语义交互,没有建立方面词与文本词的语法信息交互,导致方面词错误地关注到与其语法无关的文本词信息。此外,单词的位置距离特征和语法距离特征,分别体现其在句子线性形式中和句子语法依存树中的位置关系,而基于图卷积网络处理语法信息的方法却忽略距离特征,使距方面词较远的无关信息对其情感分析造成干扰。针对上述问题,该文提出多交互图卷积网络(MIGCN),首先将文本词位置距离特征馈入到每层图卷积网络,同时利用依存树中文本词的语法距离特征对图卷积网络的邻接矩阵加权,最后,设计语义交互和语法交互分别处理单词之间语义和语法信息。实验结果表明,在公共数据集上,准确率和宏F1值均优于基准模型。     

NgramPOS: a bigram-based linguistic and statistical feature process model for unstructured text classification

Research in financial domain has shown that sentiment aspects of stock news have a profound impact on volume trades, volatility, stock prices and firm earnings. In-depth analysis of stock news is now sourced from financial reviews by various social networking and marketing sites to help improve decision making. Nonetheless, such reviews are in the form of unstructured text, which requires natural language processing (NLP) in order to extract the sentiments. Accordingly, in this study we investigate the use of NLP tasks in effort to improve the performance of sentiment classification in evaluating the information content of financial news as an instrument in investment decision support system. At present, feature extraction approach is mainly based on the occurrence frequency of words. Therefore low-frequency linguistic features that could be critical in sentiment classification are typically ignored. In this research, we attempt to improve current sentiment analysis approaches for financial news classification by focusing on low-frequency but informative linguistic expressions. Our proposed combination of low and high-frequency linguistic expressions contributes a novel set of features for sentiment classification. The experimental results show that an optimal Ngram feature selection (combination of optimal unigram and bigram features) enhances sentiment classification accuracy as compared to other types of feature sets.

     

Sentiment classification using the theory of ANNs

Sentiment classification has attracted increasing interest from natural language processing. The goal of sentiment classification is to automatically identify whether a given piece of text expresses positive or negative opinion on a topic of interest. This paper presents the standpoint that uses individual model (i-model) based on artificial neural networks (ANNs) to determine text sentiment classification. The individual model consists of sentimental features, feature weight and prior knowledge base. During the training process, i-model that makes right sentimental judgment will correct those are wrong, to make more accurate prediction of text sentiment polarity. Experimental results show that the accuracy of individual model is higher than that of support vector machines (SVMs) and hidden Markov model (HMM) classifiers on movie review corpus.     

Sentiment analysis on Twitter: A text mining approach to the Syrian refugee crisis

The use of social media has become an integral part of daily routine in modern society. Social media portals offer powerful public platforms where people can freely share their opinions and feelings about various topics with large crowds. In the current study, we investigated the public opinions and sentiments towards the Syrian refugee crisis, which has affected millions of people and has become a widely discussed, polarizing topic in social media around the world. To analyze public sentiments about the topic on Twitter, we collected a total of 2381,297 relevant tweets in two languages including Turkish and English. Turkish sentiments were considered important as Turkey has welcomed the largest number of Syrian refugees and Turkish tweets carried information to reflect public perception of a refugee hosting country first handedly. We performed a comparative sentiment analysis of retrieved tweets. The results indicated that the sentiments in Turkish tweets were significantly different from the sentiments in English tweets. We found that Turkish tweets carried slightly more positive sentiments towards Syrians and refugees than neutral and negative sentiments, nevertheless the sentiments of tweets were almost evenly distributed among the three major categories. On the other hand, the largest number of English tweets by a significant margin contained neutral sentiments, which was followed by the negative sentiments. In comparison to the ratio of positive sentiments in Turkish tweets, 35% of all Turkish tweets, the proportion of English tweets contained remarkably less positive sentiments towards Syrians and refugees, only 12% of all English tweets.     

基于级联模型的中文情感要素抽取

随着社交媒体的发展及成熟,每天在互联网环境中都会产生大量的用户评论信息。抽取评价短语、评价对象和观点持有者等情感要素,已经成为了中文观点挖掘和情感分析的重要先决任务。针对中文情感要素抽取任务,本文提出了一个统计和规则相结合的级联模型,主要贡献包括：（1）针对汽车领域评论信息,构建情感要素标注语料库和相关词典;（2）对于以往研究较少关注的中文评价短语,本文详细分析阐述其定义和分类;（3）结合统计和规则,分别针对评价短语和情感要素提出级联抽取策略。实验结果充分证明了该级联模型的有效性,相比较于其它基于规则的情感要素抽取算法有效提升了召回率,同时为后续社交媒体情感分析任务提供了有力的支持。     

基于情感标签的极性分类

情感极性分析是文本挖掘中一种非常重要的技术.然而在不同领域中,很多情感极性分类系统存在分类精度低和缺少大量标注数据的缺陷.针对这些问题,提出了一种基于情感标签的极性分类方法.首先通过所有文本建立Sentiment-Topic模型,抽取出文本的情感标签;然后利用情感标签将文本划分为两个子文本,并通过Co-training算法对子文本进行分类;最后合并两个子文本的分类结果,并确定文本的情感极性.实验结果表明该方法具有较高的分类精度,而且不需要大量的分类样本.     

基于知识迁移和注意力融合的方面级文本情感分析

方面级情感分析是针对一个评论中涉及多种方面类别时的情感分析,现有方法通常利用方面级数据集在神经网络模型上直接进行训练,但已标注的方面级训练数据规模较小,造成模型不能充分学习而性能受限。为解决上述问题,本文利用迁移学习的思想,将数据量较大的文档级数据进行情感分析模型的预训练,进而获得丰富的文本语义、句法信息和情感特征,然后通过本文设计的目标函数及注意力融合方法,将文档级情感分析模型中的注意力权重融合到方面级情感分析模型中,从而使方面级文本情感分析性能提升。将该模型在SemEval2014数据集上进行实验,实验结果中的准确率和F1值均高于对比模型,证明了本文模型的有效性。      

基于异质网络的意见领袖社区发现

针对意见领袖社区发现问题,通过将论坛中主题及其回复关系建模为异质网络,准确表示社区结构.提出意见领袖社区影响力概念及其量化方法,在此基础上设计了一种基于异质网络的意见领袖社区发现算法.通过采集天涯论坛的大量数据,验证了该社区挖掘方案能够较准确地挖掘论坛中的意见领袖社区.