Microblog Text Research Articles

A Method for Identifying the Mood States of Social Network Users Based on Cyber Psychometrics

Analyzing people’s opinions, attitudes, sentiments, and emotions based on user-generated content (UGC) is feasible for identifying the psychological characteristics of social network users. However, most studies focus on identifying the sentiments carried in the micro-blogging text and there is no ideal calculation method for users’ real emotional states. In this study, the Profile of Mood State (POMS) is used to characterize users’ real mood states and a regression model is built based on cyber psychometrics and a multitask method. Features of users’ online behavior are selected through structured statistics and unstructured text. Results of the correlation analysis of different features demonstrate that users’ real mood states are not only characterized by the messages expressed through texts, but also correlate with statistical features of online behavior. The sentiment-related features in different timespans indicate different correlations with the real mood state. The comparison among various regression algorithms suggests that the multitask learning method outperforms other algorithms in root-mean-square error and error ratio. Therefore, this cyber psychometrics method based on multitask learning that integrates structural features and temporal emotional information could effectively obtain users’ real mood states and could be applied in further psychological measurements and predictions.

Multi-kernel SVM based depression recognition using social media data

Depression has become the world’s fourth major disease. Compared with the high incidence, however, the rate of depression medical treatment is very low because of the difficulty of diagnosis of mental problems. The social media opens one window to evaluate the users’ mental status. With the rapid development of Internet, people are accustomed to express their thoughts and feelings through social media. Thus social media provides a new way to find out the potential depressed people. In this paper, we propose a multi-kernel SVM based model to recognize the depressed people. Three categories of features, user microblog text, user profile and user behaviors, are extracted from their social media to describe users’ situations. According to the new characteristics of social media language, we build a special emotional dictionary consisted of text emotional dictionary and emoticon dictionary to extract microblog text features for word frequency statistics. Considering the heterogeneity between text feature and another two features, we employ multi-kernel SVM methods to adaptively select the optimal kernel for different features to find out users who may suffer from depression. Compared with Naive Bayes, Decision Trees, KNN, single-kernel SVM and ensemble method (libD3C), whose error reduction rates are 38, 43, 22, 21 and 11% respectively, the error rate of multi-kernel SVM method for identifying the depressed people is reduced to 16.54%. This indicates that the multi-kernel SVM method is the most appropriate way to find out depressed people based on social media data.

How will text size influence the length of its linguistic constituents?

AbstractAs human language is a multi-level complex adaptive system, a text can be seen as emergent from the complex interactions between internal and external factors. Text types such as microblog have size restriction, an external variable which may affect relevant quantitative properties of the texts themselves. Such texts provide a good opportunity to investigate the interactions between the external and internal factors of human language from the perspective of complex adaptive system. This study focuses on how the size restriction of Chinese microblog texts affects the length of their sentences and clauses. Quantitative properties concerning sentence and clause length of Chinese microblog texts are analyzed and compared with those of texts with no size restriction (i.e., prose, news report and romantic fiction). Analysis of sentence length distribution shows that size restriction has an impact on sentence length measured by numbers of words and clauses. The correlation between sentence length and clause length is examined with Menzerath-Altmann law. The satisfying fitting result of microblog texts probably suggests that, language units are able to re-organize and re-regulate their structures according to relevant external factors, until they reach a balanced new system.

Extracting Personae Interactive Relation in Chinese Microblog Based on an Improved Dependency Trigram Kernel

As the rapid development of social media, microblog has arrested a lot of attention. However, the automatic information extraction task of the microblog is relatively rare because the microblog text is quite complex and irregular. In this paper, we utilized dependency trigram kernel to construct persons relations extraction model. Firstly, we described the dependency trigram kernel (DTK) for relations extraction. Secondly, we used words semantic similarity tool HowNet to improve semantic similarity of dependency trigram's words. And then we proposed "(POS,GR) (Part of Speech,Grammatical role)" pair to improve dependency trigram's words syntax similarity. Finally, we evaluated the validity of the relation extraction model by experimenting, and the results of experiment show that the F-value of our improved DTK is higher than original DTK for microblog persons relation extraction.

A Hot Topic Identification Based on LDA and K-means++ Algorithm for Tibetan Microblog

as microblog grows more popular, services like Sina Weibo have become information providers on a web scale. Tibetan microblog is one of the most popular Tibetan network media. Researches on Tibetan microblog are now increasing. However, because of the special features of microblog text and the features of Tibetan language, traditional hot topic method of microblog cannot satisfy the need. This paper proposes a method that is a hot topic identification based on LDA and K-means++ algorithm for Tibetan microblog. Firstly, we use LDA to model microblog corpus to determine the best number of themes by the perplexity, and achieve parameters estimation with Gibbs sampling algorithm, then we can get the probability distribution of the theme and the word and the probability distribution of the theme and the document. Secondly, this paper use K-means++to cluster Tibetan microblog which have same or similar theme. Thirdly, we calculate and identify hot topics and hot words. Experimental results show that, compared with SVM method and LSA+K-means++ method, the accuracy of this method of LDA+K-means++ is higher.

Mining microblog user interests based on TextRank with TF-IDF factor

It is of great value and significance to model the interests of microblog user in terms of business and sociology. This paper presents a framework for mining and analyzing personal interests from microblog text with a new algorithm which integrates term frequency-inverse document frequency (TF-IDF) with TextRank. Firstly, we build a three-tier category system of user interest based on Wikipedia. In order to obtain the keywords of interest, we preprocess the posts, comments and reposts in different categories to select the keywords which appear both in the category system and microblogs. We then assign weight to each category and calculate the weight of keyword to get TF-IDF factors. Finally we score the ranking of each keyword by the TextRank algorithm with TF-IDF factors. Experiments on real Sina microblog data demonstrate that the precision of our approach significantly outperforms other existing methods.

MR-LDA

Latent Dirichlet Allocation(LDA) is an efficient method of text mining,but applying LDA directly to Chinese micro-blog texts will not work well because micro-blogs are more social, brief, and closely related with each other. Based on LDA, this paper proposes a Micro-blog Relation LDA model (MR-LDA), which takes the relations between Chinese micro-blog documents and other Chinese micro-blog documents into consideration to help topic mining in micro-blog. The authors extend LDA in the following two points. First, they aggregate several Chinese micro-blogs as a single micro-blog document to solve the problem of short texts. Second, they model the generation process of Chinese micro-blogs more accurately by taking relationship between micro-blog documents into consideration. MR-LDA is more suitable to model Chinese micro-blog data. Gibbs sampling method is borrowed to inference the model. Experimental results on actual datasets show that MR-LDA model can offer an effective solution to text mining for Chinese micro-blog.

Building associated semantic representation model for the ultra-short microblog text jumping in big data

In the massive microblog texts, the ultra-short microblog text is difficult to be independently understood because of its special characteristics such as data sparseness, content fragmentation and ...

Microblog Dimensionality Reduction—A Deep Learning Approach

Exploring potentially useful information from huge amount of textual data produced by microblogging services has attracted much attention in recent years. An important preprocessing step of microblog text mining is to convert natural language texts into proper numerical representations. Due to the short-length characteristics of microblog texts, using term frequency vectors to represent microblog texts will cause “sparse data” problem. Finding proper representations of microblog texts is a challenging issue. In this paper, we apply deep networks to map the high-dimensional representations of microblog texts to low-dimensional representations. To improve the result of dimensionality reduction, we take advantage of the semantic similarity derived from two types of microblogspecific information, namely the retweet relationship and hashtags. Two types of approaches, including modifying training data and modifying the training objective of deep networks, are proposed to make use of microblog-specific information. Experiment results show that the deep models perform better than traditional dimensionality reduction methods such as latent semantic analysis and latent Dirichlet allocation topic model, and the use of microblog-specific information can help to learn better representations.

Micro-blog topic detection method based on BTM topic model and K-means clustering algorithm

The development of micro-blog, generating large-scale short texts, provides people with convenient communication. In the meantime, discovering topics from short texts genuinely becomes an intractable problem. It was hard for traditional topic model-to-model short texts, such as probabilistic latent semantic analysis (PLSA) and Latent Dirichlet Allocation (LDA). They suffered from the severe data sparsity when disposed short texts. Moreover, K-means clustering algorithm can make topics discriminative when datasets is intensive and the difference among topic documents is distinct. In this paper, BTM topic model is employed to process short texts–micro-blog data for alleviating the problem of sparsity. At the same time, we integrating K-means clustering algorithm into BTM (Biterm Topic Model) for topics discovery further. The results of experiments on Sina micro-blog short text collections demonstrate that our method can discover topics effectively.

CMiner: Opinion Extraction and Summarization for Chinese Microblogs

Sentiment analysis of microblog texts has drawn lots of attention in both the academic and industrial fields. However, most of the current work only focuses on polarity classification. In this paper, we present an opinion mining system for Chinese microblogs called CMiner. Instead of polarity classification, CMiner focuses on more complicated opinion mining tasks - opinion target extraction and opinion summarization. Novel algorithms are developed for the two tasks and integrated into the end-to-end system. CMiner can help to effectively understand the users' opinion towards different opinion targets in a microblog topic. Specially, we develop an unsupervised label propagation algorithm for opinion target extraction. The opinion targets of all messages in a topic are collectively extracted based on the assumption that similar messages may focus on similar opinion targets. In addition, we build an aspect-based opinion summarization framework for microblog topics. After getting the opinion targets of all the microblog messages in a topic, we cluster the opinion targets into several groups and extract representative targets and summaries for each group. A co-ranking algorithm is proposed to rank both the opinion targets and microblog sentences simultaneously. Experimental results on a benchmark dataset show the effectiveness of our system and the algorithms.

Sentiment analysis for Chinese microblog based on deep neural networks with convolutional extension features

Related research for sentiment analysis on Chinese microblog is aiming at the analysis procedure of posts. The length of short microblog text limits feature extraction of microblog. Tweeting is the process of communication with friends, so that microblog comments are important reference information for related post. A contents extension framework is proposed in this paper combining posts and related comments into a microblog conversation for features extraction. A novel convolutional auto encoder is adopted which can extract contextual information from microblog conversation as features for the post. A customized DNN (Deep Neural Network) model, which is stacked with several layers of RBM (Restricted Boltzmann Machine), is implemented to initialize the structure of neural network. The RBM layers can take probability distribution samples of input data to learn hidden structures for better high level features representation. A ClassRBM (Classification RBM) layer, which is stacked on top of RBM layers, is adopted to achieve the final sentiment classification label for the post. Experimental results show that, with proper structure and parameters, the performance of proposed DNN on sentiment classification is better than state-of-the-art surface learning models such as SVM or NB, which proves that the proposed DNN model is suitable for short-length document classification with the proposed feature dimensionality extension method.

Mining Parallel Corpora from Sina Weibo and Twitter

Microblogs such as Twitter, Facebook, and Sina Weibo (China's equivalent of Twitter) are a remarkable linguistic resource. In contrast to content from edited genres such as newswire, microblogs contain discussions of virtually every topic by numerous individuals in different languages and dialects and in different styles. In this work, we show that some microblog users post “self-translated” messages targeting audiences who speak different languages, either by writing the same message in multiple languages or by retweeting translations of their original posts in a second language. We introduce a method for finding and extracting this naturally occurring parallel data. Identifying the parallel content requires solving an alignment problem, and we give an optimally efficient dynamic programming algorithm for this. Using our method, we extract nearly 3M Chinese–English parallel segments from Sina Weibo using a targeted crawl of Weibo users who post in multiple languages. Additionally, from a random sample of Twitter, we obtain substantial amounts of parallel data in multiple language pairs. Evaluation is performed by assessing the accuracy of our extraction approach relative to a manual annotation as well as in terms of utility as training data for a Chinese–English machine translation system. Relative to traditional parallel data resources, the automatically extracted parallel data yield substantial translation quality improvements in translating microblog text and modest improvements in translating edited news content.

Forecasting smog-related health hazard based on social media and physical sensor

Smog disasters are becoming more and more frequent and may cause severe consequences on the environment and public health, especially in urban areas. Social media as a real-time urban data source has become an increasingly effective channel to observe people׳s reactions on smog-related health hazard. It can be used to capture possible smog-related public health disasters in its early stage. We then propose a predictive analytic approach that utilizes both social media and physical sensor data to forecast the next day smog-related health hazard. First, we model smog-related health hazards and smog severity through mining raw microblogging text and network information diffusion data. Second, we developed an artificial neural network (ANN)-based model to forecast smog-related health hazard with the current health hazard and smog severity observations. We evaluate the performance of the approach with other alternative machine learning methods. To the best of our knowledge, we are the first to integrate social media and physical sensor data for smog-related health hazard forecasting. The empirical findings can help researchers to better understand the non-linear relationships between the current smog observations and the next day health hazard. In addition, this forecasting approach can provide decision support for smog-related health hazard management through functions like early warning.

Gender classification of microblog text based on authorial style

Gender profiling of unstructured text data has several applications in areas such as marketing, advertising, legal investigation, and recommender systems. The automatic detection of gender in microblogs, like twitter, is a difficult task. It requires a system that can use knowledge to interpret the linguistic styles being used by the genders. In this paper, we try to provide this knowledge for such a system by considering different sets of features, which are relatively independent of the text, such as function words and part of speech n-grams. We test a range of different feature sets using two different classifiers; namely Naive Bayes and maximum entropy algorithms. Our results show that the gender detection task benefits from the inclusion of features that capture the authorial style of the microblog authors. We achieve an accuracy of approximately 71 %, which outperforms the classification accuracy of commercially available gender detection software like Gender Genie and Gender Guesser.

Geocoding of Spatial Relationships Contained in Tweets

Human ability to understand approximate references to locations, disambiguated by means of context and reasoning about spatial relationships, is the key to describe spatial environments and to share information about them. In this paper, the authors propose an approach for geocoding that takes advantage of the spatial relationships contained in the text of tweets, using semantic and spatial analyses. Microblog text has special characteristics (e.g. slang, abbreviations, acronyms, etc.) and thus represents a special variation of natural language. The main objective of this work is to associate spatial relationships found in text with a spatial footprint, to determine the location of the event described in the tweet. The feasibility of the proposal is demostrated using a corpus of 200,000 tweets posted in Spanish related with traffic events in Mexico City.

Modeling information diffusion on microblog networks based on users' behaviors

Online social networks, such as Facebook, Twitter and YouTube, play a vital role in information sharing and diffusion, and recently many dynamics models on social networks have been proposed to model information diffusion. However most models are theoretical, their parameters do not come from realistic data and their validity and reliability have not been evaluated empirically. In the paper we first analyze the users' behaviors of reading and reposting microblog in Sina Weibo, a Twitter-like website in China, and find that users' number of fans, the average reposted number of users' microblog, the intensity of users' interaction and the similarity between microblog topics and users' topic interests can significantly influence reposting behavior. Then we propose an information diffusion model Susceptible-Infected-Recovered based on Users' Behaviors (SIRUB) on microblog networks, compute the users' probability of reading microblog in the model according to the probability of their logging on microblog in a day, and obtain the reposting probability utilizing the logistic regression which considers 16 possible factors influencing users' reposting behavior. The 16 factors can be divided into three categories: the characteristics of microblog publishers, microblog text features and social relationship characteristics. We utilize the beginning 2/3 microblog data to obtain model parameters and logistic regression coefficients, and the remaining 1/3 data to examine the validity of the model. The experiments on Sina Weibo network show that the model can predict users' reposting behavior accurately only when it considers both reading and reposting probabilities. F-score which considers precision and recall is used to assess prediction effect of the model. The highest F-score for the prediction of SIRUB model on users' reposting behavior is 0.228 which is much larger than those of classical Susceptible-Infected-Recovered (SIR, F-score=0.039) and Susceptible-Infected-Contacted-Recovered (SICR, F-score=0.037) models. The prediction on the spreading scope of microblog for SIR and SICR models is related with users' number of fans while for SIRUB model not. For SIRUB model the mean and standard deviation of the errors of prediction on spreading scope are smaller than those of SIR and SICR models. These results indicate that users' behaviors of reading and reposting microblog should be appropriately taken in account when modeling information diffusion on microblog networks, and that, in general, the prediction performance of the data-driven SIRUB model proposed in the paper is better than those of SIR and SICR models regardless of the prediction of users' reposting behavior or diffusion scope of microblog.

Opinion mining of microblog texts on Hadoop ecosystem

Opinion mining started getting more traction due to the increasing popularity of Twitter and similar social network platforms that are producing fast and real-time responses to social events. It is a very challenging area since it is difficult, if not impossible, to identify general public sentiment towards events, entities, etc., using opinion mining techniques over huge numbers of tweets and messages automatically. In this study we present our opinion mining techniques on tweet data with early results. We apply sentiment scoring and clustering algorithms using Hadoop ecosystem for parallel processing. We classify tweets by tagging them as positive, negative, and neutral as a result.

ИССЛЕДОВАНИЕ ОСОБЕННОСТЕЙ ПОНИМАНИЯ ТЕКСТОВ МИКРОБЛОГА (на материале текста-комментария)

ИССЛЕДОВАНИЕ ОСОБЕННОСТЕЙ ПОНИМАНИЯ ТЕКСТОВ МИКРОБЛОГА (на материале текста-комментария)

Structure in the Tweet Haystack: Uncovering the Link between Text-Based Sentiment Signals and Financial Markets

We examine the relationship between signals derived from unstructured social media microblog text data and financial market developments. Employing statistical language modeling techniques we construct directional user sentiment and non-directional topic disagreement metrics and link these to S&P 500 index returns and volatility. Based on an extensive five year sample of Twitter messages our study shows that both unsupervised and supervised statistical learning methods successfully identify subsets of expert users with distinct finance focus. This allows to filter out the substantial noise associated with social media. Accounting for salient properties of the time series in ARMA models we document significant effects of expert disagreement signals on current and future S&P volatility. Moreover, we detect a significant contemporaneous relation between expert sentiment signals and S&P returns.