Bayesian method infers ages of 700M Twitter users from follow data.
problem Lack of demographic data on Twitter users.
method Bayesian framework to generalize age information from a few users to the entire network.
result Highly accurate age inference for 700 million Twitter accounts.
The paper uses NMF to detect political communities in Twitter networks.
problem Detecting pure political communities in Twitter networks.
method Developed three NMF frameworks to analyze user connectivity and content.
result User content and endorsement filtered connectivity are complementary.
A novel deep learning method predicts Twitter users' locations using multiple data types.
problem Predicting Twitter users' locations on large social networks.
method Combines content-based and network-based approaches using a multi-entry neural network architecture (MENET).
result MENET outperforms state-of-the-art methods by a large margin on three benchmark datasets.
State-of-the-art Twitter sentiment analysis using CNNs and LSTMs.
problem Improving Twitter sentiment classification accuracy.
method Pre-trained word embeddings, distant supervision, fine-tuning, ensemble of CNNs and LSTMs.
result First rank on all five English subtasks of SemEval-2017.
Deep neural networks detect cyberthreats from Twitter.
problem Scalable processing of cybersecurity information from social media.
method Convolutional and bidirectional LSTM networks for tweet classification and named entity recognition.
result Pipeline achieves high true positive and F1-score rates for threat detection.
Twitter releases dataset to study user engagement on Home Timeline, focusing on privacy.
problem Limited exploration of user feedback on Twitter's Home Timeline.
method Release of a large-scale dataset of 160 million Tweets with engagement information.
result Demonstrates challenges in predicting user engagements and privacy concerns.
Study shows how Twitter users' sentiment levels correlate with network structure and external events.
problem Understanding how Twitter users' sentiment levels are influenced by network structure and external events.
method Used a large dataset of tweets, applied sentiment scoring algorithms, and created an agent-based model.
result People with larger communication reach use positive sentiment more often and negative sentiment less often.
Paper introduces score embedding for Twitter sentiment analysis of health care issues.
problem Analyzing public opinion on health care plans using social media.
method Score embedding, a neural network model for word representations.
result Score embedding effectively captures sentiment and outperforms existing methods.
New metric improves social message classification on Twitter.
problem Efficient classification of short social messages on online networks.
method Dynamic Time Warping distance for message spreading classification.
result Good classification accuracies on Twitter data.
Study uses deep learning to predict gender and analyze HPV vaccine perceptions on Twitter.
problem Analyzing gender differences in public perceptions on HPV vaccine using social media data.
method Convolutional neural network model trained on Twitter text for gender prediction, then applied to HPV vaccine related tweets.
result Identified gender differences in public perceptions on HPV vaccine, consistent with previous studies.
Crowdsourced algorithms identify fake news on Twitter.
problem Identifying fake news on social media platforms.
method Evaluation of reputation algorithms on a large dataset of Twitter news.
result Simple crowdsourcing-based algorithms can identify a significant portion of fake news with low false positive rates.
Study uses machine learning to detect depression in Twitter users.
problem Detecting depression in Twitter users using social media data.
method Machine learning techniques applied to Twitter activity and tweet features.
result More features improve accuracy in detecting depressed users.
It has often been taken as a working assumption that directed links in information networks are frequently formed by "short-cutting" a two-step path between the source and the destination -- a kind of implicit "link copying" analogous to the process of triadic closure in social networks. Despite the role of this assump…
Study shows frequent 'stock' mentions on Twitter correlate with stock market declines.
problem Understanding how social media impacts financial markets during pandemics.
method Collected over 26 million tweets about COVID-19, analyzed using natural language processing.
result More frequent use of 'stock' in tweets predicts substantial declines in stock market indices.
This paper tackles spam detection on Twitter by analyzing correlated features.
problem Spam detection on social media, especially Twitter, to improve user experience.
method Extracted tweet-based and user-based features, identified correlated features, and used artificial neural networks for classification.
result Achieved 97.57% accuracy in classifying tweets as spam or non-spam.
Simplified LSTM models improve sentiment analysis on Twitter debate data.
problem Performing sentiment analysis on long sequence data from Twitter debates.
method Developed six parameter-reduced LSTM models (slim LSTM) for faster training and reduced computational cost.
result Slim LSTM models outperform standard LSTM model in sentiment analysis of GOP Debate Twitter dataset.
TIMME detects Twitter users' ideology from sparse, heterogeneous data.
problem Predicting political tendencies from Twitter data.
method TIMME, a multi-task multi-relational embedding model.
result TIMME outperforms state-of-the-art models for ideology detection.
This paper examines feature selection for extracting user intentions from Twitter.
problem Extracting user intentions from informal, misspelled tweets.
method Developed a dataset from Twitter feeds, used two feature selection techniques (Information Gain and hybrid forward selection), and applied four classification algorithms.
result The hybrid feature selection approach outperformed the Information Gain method.
Study shows how socioeconomic status influences language use on Twitter.
problem Global variability of linguistic patterns due to socioeconomic factors.
method Multivariate analysis of French Twitter corpus and socioeconomic data.
result People with higher socioeconomic status use more standard language.
Study analyzes tweets from Northeastern libraries to understand public relations efforts.
problem Analyzing vast social media data for public relations.
method Computational methods to extract and analyze tweets.
result Libraries use Twitter to represent services and disseminate information.
This study uses Twitter to analyze traveler behavior in Manhattan.
problem Analyzing traveler behavior using social media data.
method Systematic method to extract displacement information from geo-tagged tweets.
result Twitter reveals unique demographics and travel behavior patterns.
Twitter promotes cryptocurrency pump-and-dumps, affecting trading behavior and returns.
problem The influence of Twitter on cryptocurrency pump-and-dump events.
method Analysis of abnormal returns, trading volume, and tweet activity.
result Investors relying on Twitter information sell later, leading to significant losses.
This paper classifies tweets into positive and negative sentiments using distributed word and sentence representations.
problem Classifying tweets into positive and negative sentiments.
method Used distributed representations of words and sentences, and LSTM and CNN networks for classification.
result Achieved accuracies as high as 81%.
Predicts Bitcoin price using Twitter sentiment analysis.
problem Volatility and varied opinions in cryptocurrency markets.
method Developed a model combining sentiment analysis of tweets and historical price data.
result Sentiment prediction MAPE of 9.45%, price prediction MAPE of 3.6%
Study explores rumor spread on Twitter using supervised learning.
problem Rumors spread quickly on social networks, affecting decision-making.
method Microscopic-level analysis of user interactions and features.
result Models predict rumor diffusion with 90% accuracy.
Framework detects influential actors in disinformation networks.
problem Identifying and countering hostile influence operations on social media.
method Combines NLP, ML, graph analytics, and causal inference.
result 96% precision, 79% recall, 96% PR-curve area for IO detection.
Study validates Twitter search protocol for health and disease events.
problem Validating search protocols for health-related tweets on Twitter.
method Extracted and analyzed health and disease tweets using predefined keywords, trained CART algorithms for predictions.
result Validation showed good accuracy (AUC > 0.8) in detecting relevant health and disease tweets.
The study analyzes sentiment of European tweets during the pandemic.
problem Understanding public sentiment during the COVID-19 pandemic.
method Cross-language sentiment analysis of multilingual tweets using neural networks and sentence embeddings.
result Sentiment analysis reveals that lockdown announcements correlate with a deterioration of mood, which recovers quickly.
The study builds a classifier to filter trustworthy tweets from rumours on Twitter.
problem Misinformation on social media, specifically rumours on Twitter.
method Collected 100 million tweets, identified 72 rumours, used 80 measures for trustworthiness, trained machine learning classifiers.
result The model was more accurate than previous studies, identifying key attributes of trustworthy tweets.
Novel TM-vector model predicts stock market direction using Twitter and market data.
problem Challenging stock market forecasting with equal or ignored user effects.
method TM-vector trained with Twitter features and market information, using IndRNN.
result Significant accuracy in predicting stock market direction, especially for Apple.
This study analyzes Twitter posts from US west coast libraries.
problem Lack of automated methods to analyze thousands of library tweets.
method Used Twitter API to collect and analyze 138,000 tweets.
result Identified 20 topics in library tweets, categorized into public relations, book, event, training, and social good.
Research shows Twitter is permeable to financial events, influencing its content and sentiment.
problem Investigating how Twitter reacts to financial events.
method Conducted experiments on a specific financial event (Tesco PLC and Booker Group PLC merger announcement).
result Twitter is permeable to financial events, affecting its content and sentiment.
Study shows Twitter sentiments predict stock price fluctuations.
problem Predicting stock prices using public opinions.
method Time series analysis and natural language processing with LSTM model.
result Positive, negative, and subjective sentiments correlate with stock price changes.
Kernel analysis reveals rumor truth from diffusion patterns alone.
problem Detecting unverified rumors on Twitter using text and user identities.
method Graph kernels to extract diffusion patterns from Twitter cascade structures.
result Diffusion patterns are highly informative of rumor truth or falsehood.
Analyzes Twitter users' opinions on self-driving cars.
problem Understanding public perception of self-driving cars.
method Annotated Twitter dataset, topic modeling, sentiment classification using Twitter features.
result People are generally optimistic but also concerned about self-driving cars.
Twitter, a popular social network, presents great opportunities for on-line machine learning research. However, previous research has focused almost entirely on learning from passively collected data. We study the problem of learning to acquire followers through normative user behavior, as opposed to the mass following…
A new Twitter sentiment model predicts stock market trends with high accuracy.
problem Real-time prediction of future stock market prices.
method Baseline correlation approach using polynomial regression, classification, and lexicon-based sentiment analysis.
result Predicts stock market trends with 67.22% accuracy, up to 15 time samples in advance.
Twitter has been proven to be a notable source for predictive modelling on various domains such as the stock market, the dissemination of diseases or sports outcomes. However, such a study has not been conducted in football (soccer) so far. The purpose of this research was to study whether data mined from Twitter can b…
For classifying time series, a nearest-neighbor approach is widely used in practice with performance often competitive with or better than more elaborate methods such as neural networks, decision trees, and support vector machines. We develop theoretical justification for the effectiveness of nearest-neighbor-like clas…
Model predicts stock prices using Twitter sentiment data.
problem Predicting stock prices using social media sentiment.
method Extracted 19 sentiment features from Twitter posts about Petrobras, trained XBoot models, and simulated trading.
result Simulated trading gained R$88,82 (net) over 250 days.
Amobee's system won 3rd place in Twitter sentiment classification.
problem Sentiment detection on Twitter using deep learning.
method RNN models trained on a sentiment treebank, combined with classifiers.
result 3rd place in SemEval 2017 task 4, 5-label classification.
New model detects influential users in hidden social networks.
problem Detecting influential users in networks with unknown structure.
method Developed a multi-task low rank linear influence model.
result Successfully identifies influential users for specific topics.
Twitter data reveals global movement patterns for social good.
problem Lack of accessible, unbiased, high-resolution spatial data.
method Used geolocation data from Twitter to estimate global mobility patterns.
result Demonstrated Twitter data can address spatial data shortcomings.
Predicting depression on Twitter using social media posts.
problem Identifying depression in online personas.
method Crowdsourced Twitter users, Bag of Words approach, statistical classifiers.
result 81% accuracy rate in depression risk estimation.
Twitter system detects unexpected epidemics.
problem Detecting sudden, unexpected epidemic outbreaks in Twitter.
method Dynamic classification, alert generation, and ranking/recommendation.
result Empirical evaluation and validation with domain experts.
Study maps Spanish dialects on Twitter, revealing urban vs regional variations.
problem Understanding linguistic diversity across Spanish-speaking regions.
method Geographically tagged Twitter corpus, machine learning for dialect analysis.
result Urban dialects show international characteristics, rural dialects regional uniformity.
Algorithm uses Twitter sentiment analysis for short-term stock trading.
problem Automated short-term stock trading on the stock market.
method Adaptive algorithm using semantic sentiment analysis of Twitter posts.
result Our approach outperforms a well-known benchmark and achieves near-optimal performance.
ED-Filter improves eating disorder classification on Twitter.
problem High dimensionality and extensive feature sets in Twitter data for ED classification.
method Informed branch and bound search technique with hybrid greedy-based deep learning.
result Significant improvements in classification accuracy and efficiency.