Financial Tweets Research Articles

Discretionary dissemination on Twitter

AbstractThe study provides large‐scale descriptive evidence on the timing and nature of corporate financial tweeting. Using an unsupervised machine learning approach to analyze 24 million tweets posted by S&P 1500 firms from 2012 to 2020, we find that firms are more likely to tweet financial information around significantly negative or positive news events, such as earnings announcements and the filing of financial statements. This convex U‐shaped relation between the likelihood of posting financial tweets and the materiality of accounting events becomes stronger over time. Whereas research based on early samples concludes that firms are less likely to disseminate financial information on Twitter when the news is bad and material, the symmetric dissemination behavior we find suggests that these conclusions should be revised. We also show that a machine learning algorithm (Twitter‐Latent Dirichlet Allocation) is superior to a dictionary approach in classifying short messages like tweets.

Prediction of Machine-Generated Financial Tweets Using Advanced Bidirectional Encoder Representations from Transformers

With the rise of Large Language Models (LLMs), distinguishing between genuine and AI-generated content, particularly in finance, has become challenging. Previous studies have focused on binary identification of ChatGPT-generated content, overlooking other AI tools used for text regeneration. This study addresses this gap by examining various AI-regenerated content types in the finance domain. Objective: The study aims to differentiate between human-generated financial content and AI-regenerated content, specifically focusing on ChatGPT, QuillBot, and SpinBot. It constructs a dataset comprising real text and AI-regenerated text for this purpose. Contribution: This research contributes to the field by providing a dataset that includes various types of AI-regenerated financial content. It also evaluates the performance of different models, particularly highlighting the effectiveness of the Bidirectional Encoder Representations from the Transformers Base Cased model in distinguishing between these content types. Methods: The dataset is meticulously preprocessed to ensure quality and reliability. Various models, including Bidirectional Encoder Representations Base Cased, are fine-tuned and compared with traditional machine learning models using TFIDF and Word2Vec approaches. Results: The Bidirectional Encoder Representations Base Cased model outperforms other models, achieving an accuracy, precision, recall, and F1 score of 0.73, 0.73, 0.73, and 0.72 respectively, in distinguishing between real and AI-regenerated financial content. Conclusions: This study demonstrates the effectiveness of the Bidirectional Encoder Representations base model in differentiating between human-generated financial content and AI-regenerated content. It highlights the importance of considering various AI tools in identifying synthetic content, particularly in the finance domain in Pakistan.

Comparative Study for Sentiment Analysis of Financial Tweets with Deep Learning Methods

Nowadays, Twitter is one of the most popular social networking services. People post messages called “tweets”, which may contain photos, videos, links and text. With the vast amount of interaction on Twitter, due to its popularity, analyzing Twitter data is of increasing importance. Tweets related to finance can be important indicators for decision makers if analyzed and interpreted in relation to stock market. Financial tweets containing keywords from the BIST100 index were collected and the tweets were tagged as “POSITIVE”, “NEGATIVE” and “NEUTRAL”. Binary and multi-class datasets were created. Word embedding and pre-trained word embedding were used for tweet representation. As classifiers, Neural Network, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU) and GRU-CNN models were used in this study. The best results for binary and multi-class datasets were observed with pre-trained word embedding with the CNN model (83.02%, 72.73%). When word embedding was employed, the Neural Network model had the best results on the multi-class dataset (63.85%) and GRU-CNN had the best results on the binary dataset (80.56%).

Evaluation of transformer models for financial targeted sentiment analysis in Spanish.

Nowadays, financial data from social media plays an important role to predict the stock market. However, the exponential growth of financial information and the different polarities of sentiment that other sectors or stakeholders may have on the same information has led to the need for new technologies that automatically collect and classify large volumes of information quickly and easily for each stakeholder. In this scenario, we conduct a targeted sentiment analysis that can automatically extract the main economic target from financial texts and obtain the polarity of a text towards such main economic target, other companies and society in general. To this end, we have compiled a novel corpus of financial tweets and news headlines in Spanish, constituting a valuable resource for the Spanish-focused research community. In addition, we have carried out a performance comparison of different Spanish-specific large language models, with MarIA and BETO achieving the best results. Our best result has an overall performance of 76.04%, 74.16%, and 68.07% in macro F1-score for the sentiment classification towards the main economic target, society, and other companies, respectively, and an accuracy of 69.74% for target detection. We have also evaluated the performance of multi-label classification models in this context and obtained a performance of 71.13%.

PyFin-sentiment: Towards a machine-learning-based model for deriving sentiment from financial tweets

Responding to the poor performance of generic automated sentiment analysis solutions on domain-specific texts, we collect a dataset of 10,000 tweets discussing the topics of finance and investing. We manually assign each tweet its market sentiment, i.e., the investor’s anticipation of a stock’s future return. Using this data, we show that all existing sentiment models trained on adjacent domains struggle with accurate market sentiment analysis due to the task’s specialized vocabulary. Consequently, we design, train, and deploy our own sentiment model. It outperforms all previous models (VADER, NTUSD-Fin, FinBERT, TwitterRoBERTa) when evaluated on Twitter posts. On posts from a different platform, our model performs on par with BERT-based large language models. We achieve this result at a fraction of the training and inference costs due to the model’s simple design. We publish the artifact as a python library to facilitate its use by future researchers and practitioners.

Stock Prediction Based on Twitter Sentiment Extraction Using BiLSTM-Attention

A profitable stock price prediction will yield a large profit. According to behavioural economics, other people's emotions and viewpoints have a significant impact on business. One of them is the rise and fall of stock prices. Previous studies have shown that public sentiments retrieved from online information can be very valuable on market trading. In this paper, we propose a model that works well in predicting future stock prices by using public sentiments from social media. The online information used in this research is financial tweets collected from Twitter and the stock prices values retrieved from Yahoo! Finance. We collected tweets related to Netflix Company stocks and the stock prices for the same period which is 5 years from 2015 to 2020 as the dataset. We extracted the sentiment value using VADER algorithm. In this paper, we apply a Bidirectional Long Short-Term Memory (BiLSTM) architecture to achieve our goal. Moreover, we created seven different experiments with different stock price parameters and selected sentiment values combinations and investigated the model by adding an attention layer. We experimented with two different sentiment values, tweet’s compound value and tweet’s compound value multiplied by favorites count. We considered the favorites count as one representation of public sentiments. From the seven experiments, the experiment with Bidirectional Long Short-Term Memory (BiLSTM) - attention model combined with our selected stock price parameters namely close price, open price, and using Twitter sentiment values that are multiplied with the tweet’s favorites count yields a better RMSE result of 2.482e-02 in train set and 2.981e-02 in the test set.

Разработка тонально-тематического словаря EcSentiThemeLex для анализа экономических текстов на русском языке

The main goal of the research is to develop a publicly available tonal-thematic dictionary in Russian, which allows identifying the semantic orientation of groups of economic texts, as well as determining their sentimental (tonal) characteristics. The article describes the main stages of compiling a dictionary using machine learning methods (clustering, word frequency allocation, correlogram construction) and expert evaluation of determining the tonality and expanding the dictionary by including terms from similar foreign dictionaries. The empirical base of the research included: annual reports of companies, news from ministries and the Central Bank of the Russian Federation, financial tweets of companies and RBC news articles in the area of "Economics, Finance, money and business". The compiled dictionary differs from the previous ones in the following ways: 1. it is one of the first dictionaries which can be used to rate the tone of economic and financial texts in Russian language by 5 degrees of tonality; 2. allows you to rate the tonality and content of the text by 12 economic topics (e. g., macroeconomics, monetary policy, stock and commodity markets, etc.) 3. the final version of EcSentiThemeLex dictionary is included in the software package (library) ‘rulexicon’ for the programming environment R and Python. Step-by-step examples of using the developed library in the R environment are given. It allows to evaluate the tone and thematic focus of an economic or financial text by means of a concise code. The structure of the library allows you to use the original texts for their assessment without prior lemmatization (the reduction to elementary forms).The resulting EcSentiThemeLex dictionary is included in the rulexicon software package for the R modeling environment .The tonal-thematic dictionary EcSentiThemeLex with all word forms compiled in this work will simplify the solution of applied problems of text analysis in the financial and economic sphere, and can also potentially serve as a basis for increasing the number of relevant studies in the Russian literature.

Predicting Stock Market Price Movement Using Sentiment Analysis: Evidence From Ghana

Abstract Predicting the stock market remains a challenging task due to the numerous influencing factors such as investor sentiment, firm performance, economic factors and social media sentiments. However, the profitability and economic advantage associated with accurate prediction of stock price draw the interest of academicians, economic, and financial analyst into researching in this field. Despite the improvement in stock prediction accuracy, the literature argues that prediction accuracy can be further improved beyond its current measure by looking for newer information sources particularly on the Internet. Using web news, financial tweets posted on Twitter, Google trends and forum discussions, the current study examines the association between public sentiments and the predictability of future stock price movement using Artificial Neural Network (ANN). We experimented the proposed predictive framework with stock data obtained from the Ghana Stock Exchange (GSE) between January 2010 and September 2019, and predicted the future stock value for a time window of 1 day, 7 days, 30 days, 60 days, and 90 days. We observed an accuracy of (49.4–52.95 %) based on Google trends, (55.5–60.05 %) based on Twitter, (41.52–41.77 %) based on forum post, (50.43–55.81 %) based on web news and (70.66–77.12 %) based on a combined dataset. Thus, we recorded an increase in prediction accuracy as several stock-related data sources were combined as input to our prediction model. We also established a high level of direct association between stock market behaviour and social networking sites. Therefore, based on the study outcome, we advised that stock market investors could utilise the information from web financial news, tweet, forum discussion, and Google trends to effectively perceive the future stock price movement and design effective portfolio/investment plans.

The Irruption of Cryptocurrencies Into Twitter Cashtags: A Classifying Solution

There is a consensus about the good sensing characteristics of Twitter to mine and uncover knowledge in financial markets, being considered a relevant feeder for taking decisions about buying or holding stock shares and even for detecting stock manipulation. Although Twitter hashtags allow to aggregate topic-related content, a specific mechanism for financial information also exists: Cashtag (consisting of the company ticker preceded by $) is a supporting mechanism to track financial tweets referring to a company listed in a stock market. However, according to our experiments and due to the lack of conventions in cashtags usage, the irruption of cryptocurrencies has resulted in a significant degradation on the cashtag-based aggregation of posts. Unfortunately, Twitter' users may use homonym tickers to refer to cryptocurrencies and to companies in stock markets, which means that filtering by cashtag may result on both posts referring to stock companies and cryptocurrencies. This research proposes automated classifiers to distinguish conflicting cashtags and, so, their container tweets by analyzing the distinctive features of tweets referring to stock companies and cryptocurrencies. As experiment, this paper analyses the interference between cryptocurrencies and company tickers in the London Stock Exchange (LSE), specifically, companies in the main and alternative market indices FTSE-100 and AIM-100. Heuristic-based as well as supervised classifiers are proposed and their advantages and drawbacks, including their ability to self-adapt to Twitter usage changes, are discussed. The experiment confirms a significant distortion in collected data when colliding or homonym cashtags exist, i.e., the same $ acronym to refer to company tickers and cryptocurrencies. According to our results, the distinctive features of posts including cryptocurrencies or company tickers support accurate classification of colliding tweets (homonym cashtags) and Independent Models, as the most detached classifiers from training data, have the potential to be trans-applicability (in different stock markets) while retaining performance.

The Chilling Effect of Regulation FD: Evidence from Twitter

We find evidence consistent with the notion that the broad non-exclusionary disclosure requirement of Reg-FD inhibits the adoption of new disclosure technologies. Our analyses exploit unanticipated SEC guidance (“Reg-SocMedia”) that outlined how social media channels could be used in compliance with Reg-FD. Overall, we find substantial changes in corporate Twitter usage to communicate financial information using a large hand-collected and coded sample. We document a non-transitory stock price response of 25 basis points to financial tweets following Reg-SocMedia, compared with no detectable response before. This differential price response occurs on days with no concurrent disclosures, and is stronger for firms with more Twitter followers. We use a variety of alternative specifications and placebo tests to provide assurance that the effects we document are attributable to Reg-SocMedia rather than time-series changes in disclosure policies. Collectively, our findings suggest that firms were reluctant to use Twitter to communicate value relevant financial information without SEC guidance, consistent with the concern that Reg-FD has a “chilling” effect on corporate disclosures.

Cashtag Piggybacking

Microblogs are increasingly exploited for predicting prices and traded volumes of stocks in financial markets. However, it has been demonstrated that much of the content shared in microblogging platforms is created and publicized by bots and spammers. Yet, the presence (or lack thereof) and the impact of fake stock microblogs has never been systematically investigated before. Here, we study 9M tweets related to stocks of the five main financial markets in the US. By comparing tweets with financial data from Google Finance, we highlight important characteristics of Twitter stock microblogs. More importantly, we uncover a malicious practice—referred to as cashtag piggybacking —perpetrated by coordinated groups of bots and likely aimed at promoting low-value stocks by exploiting the popularity of high-value ones. Among the findings of our study is that as much as 71% of the authors of suspicious financial tweets are classified as bots by a state-of-the-art spambot-detection algorithm. Furthermore, 37% of them were suspended by Twitter a few months after our investigation. Our results call for the adoption of spam- and bot-detection techniques in all studies and applications that exploit user-generated content for predicting the stock market.

Discretionary Disclosure on Twitter

Using a machine learning approach to analyze 12.8 million tweets posted by S&P 1500 firms from 2012 to 2016, we find that firms time financial tweets around earnings announcements, accounting filings as well as other important corporate events, and are more likely to use media (images or video) and links in those tweets. The above pattern holds for both good and bad news. Moreover, we find that feedback from Twitter users encourages future financial tweets and use of media and links. These results collectively suggest that firms make discretionary choices in timing and presentation format when disseminating information on social media and that they incorporate instantaneous feedback from Twitter users into their dissemination strategies.

Support Vector Machine for Sentiment Analysis of Nigerian Banks Financial Tweets

The rise of social media paves way for unprecedented benefits or risks to several organisations depending on how they adapt to its changes. This rise comes with a great challenge of gaining insights from these big data for effective and efficient decision making that can improve quality, profitability, productivity, competitiveness and customer satisfaction. Sentiment analysis is the field that is concerned with the classification and analysis of user generated text under defined polarities. Despite the upsurge of research in sentiment analysis in recent years, there is a dearth in literature on sentiment analysis applied to banks social media data and mostly on African datasets. Against this background, this study applied machine learning technique (support vector machine) for sentiment analysis of Nigerian banks twitter data within a 2-year period, from 1st January 2017 to 31st December 2018. After crawling and preprocessing of the data, LibSVM algorithm in WEKA was used to build the sentiment classification model based on the training data. The performance of this model was evaluated on a pre-labelled test dataset generated from the five banks. The results show that the accuracy of the classifier was 71.8367%. The precision for both the positive and negative classes was above 0.7, the recall for the negative class was 0.696 and that of the positive class was 0.741 which shows the prediction did better than chance in addition to other measures. Applying the model in predicting the sentiments of the five Nigerian banks twitter data reveals that the number of positive tweets within this period was slightly greater than the number of negative tweets. The scatter plots for the sentiments series indicated that, majority of the data falls between 0 and 100 sentiments per day, with few outliers above this range.

Twitter data models for bank risk contagion

A very important and timely area of research in finance is systemic risk modelling, which concerns the estimation of the relationships between different financial institutions, with the aim of establishing which of them are more contagious/subject to contagion. The aim of this paper is to develop a systemic risk model which, differently from existing ones, employs not only the information contained in financial market prices, but also big data coming from financial tweets. From a methodological viewpoint, we propose a new framework, based on graphical Gaussian models, that can estimate systemic risks with stochastic network models based on two different sources: financial markets and financial tweets, and suggest a way to combine them, using a Bayesian approach. From an applied viewpoint, we present the first systemic risk model based on big data, and show that such a model can help predicting the default probability of a bank, conditionally on the others.

A network analysis of financial conversations on Twitter

Assessing influence on social media is at the heart of a new trend in humanities and social sciences. Our paper is about assessing users' influence on social networks. We compare four different methods regarding tweets about financial conversations. We have built two datasets with respectively 489,000 and 280,000 financial tweets. While getting additional followers may provide insight of an increasing popularity, it does not necessarily translate into a more influential position. Our results suggest that the number of followers is only one element of what is considered to be influential. Indeed, the number of messages can be biased by what can be described as 'noisy' users. The number of retweets is a more refined proxy about influence. Finally, the betweenness centrality in the retweet network provides some privileged information to anyone following these users. We compare the performance of stocks mentioned by each type of users. From these results, such methods of assessing influence on Twitter leverage the acquisition of preferential signals in a financial context.

Big data analysis for financial risk management

A very important area of financial risk management is systemic risk modelling, which concerns the estimation of the interrelationships between financial institutions, with the aim of establishing which of them are more central and, therefore, more contagious/subject to contagion. The aim of this paper is to develop a novel systemic risk model. A model that, differently from existing ones, employs not only the information contained in financial market prices, but also big data coming from financial tweets. From a methodological viewpoint, the novelty of our paper is the estimation of systemic risk models using two different data sources: financial markets and financial tweets, and a proposal to combine them, using a Bayesian approach. From an applied viewpoint, we present the first systemic risk model based on big data, and show that such a model can shed further light on the interrelationships between financial institutions.

Is Corporate Tweeting Informative or is it Just Hype? Evidence from the SEC Social Media Regulation

Using textual analysis, I identify corporate tweets that contain financial information and study their impact on market prices. Prior to the April 2, 2013 SEC regulation allowing firms to use social media as an official news outlet, financial tweets appear to have been mostly hype in that prices do not respond to firms’ own tweeting. Following the regulation, tweeting appears to be informative – leading to a 19.5 basis points return on tweeting days that is not subsequently reversed. Moreover, firms strategically tweet negative information at times of low investor attention, such as Friday afternoons. The results suggest that firms actively use social media to communicate fundamental information to investors.

Categorical network models for systemic risk measurement

A very important area of financial risk management is systemic risk modelling, which concerns the estimation of the interrelationships between financial institutions, with the aim of establishing which of them are more central and, therefore, more contagious/subject to contagion. The aim of this paper is to develop a systemic risk model which, differently from existing ones, employs not only the information contained in financial market prices, but also big data coming from financial tweets. From a methodological viewpoint, we propose a new framework, based on categorical graphical models, that can estimate systemic risks with models based on two different sources: financial markets and financial tweets, and suggest a way to combine them, using a Bayesian approach. From an applied viewpoint, we present the first systemic risk model based on big data, and show that such a model can shed further light on the interrelationships between financial institutions. This can help predicting the level of returns of a bank, conditionally on the others, for example when a shock occurs in another bank.

How to measure the quality of financial tweets

Twitter text data may be very useful to evaluate from a different perspective financial tangibles, such as share prices, as well as intangible assets, such as company reputation. While twitter data are becoming widely available to researchers, methods aimed at selecting reliable twitter data are, to our knowledge, not yet available. To overcome this problem, and allow to employ twitter data for descriptive and predictive purposes, in this contribution we propose an effective statistical method that formalises and extends a quality index employed in the context of the evaluation of academic research, the h index, renamed T index. Our proposal will be tested on a list of twitterers described by the Financial Times as “the top financial tweeters to follow”, for the year 2013. Using our methodology we rank these twitterers and provide confidence intervals to decide whether they are significantly different. Moreover through a sentiment analysis, we employ the twitters content to estimate graphical models useful in the context of financial systemic risk. To this aim we focus on the Italian bank system and we show how listed banks are connected on the basis of tweets data.