Sentiment Metrics Research Articles

Unveiling the Nexus Between Crises, Investor Sentiment, and Volatility of Tourism-Related Stocks: Empirical Findings From Pakistan

This study investigates the predictive power of sentiment metrics concerning the volatility of tourism-related stocks listed on the Pakistan Stock Exchange (PSX) during recent crises. We employed univariate and multivariate GARCH-family econometric models to analyze daily data collected from March 2020 to April 2023. The data is categorized into two periods: A (COVID-19 until February 2022) and B (the vote of no-confidence motion until April 2023). The findings indicate that Market Sentiment (MS) plays a more significant role than individual sentiment in predicting stock price volatility and negatively impacts return volatility. During period A, the Fear Index (FI) and in B, economic policy uncertainty (EPU) were identified as factors contributing to increased fluctuation in tourism-related stocks. Additionally, MS exhibited a heightened level of bearishness compared to the COVID-19 era. The multivariate BEKK-GARCH analysis revealed that MS past volatility negatively spillover onto the current conditional volatility of PSX firms. This highlights the paramount importance of MS in understanding stock price fluctuations in the PSX. Interestingly, the technology sector exhibited lower sensitivity across the analysis. This study provides valuable insights for decision-makers by elucidating the principles of behavioral finance theory within the context of the PSX. They should prioritize long-term strategies to institutionalize investors, enhance their emotional intelligence, and mitigate psychological biases in the market.

We care: A User-Focused Sentiment Analysis Framework for Social Media Interactions

This paper addresses the sentiment analysis of WhatsApp chats, focusing on the challenges and opportunities presented by conversational data. The introduction outlines the prevalence of sentiment analysis in social media platforms and the significance of accurate sentiment categorization. The literature review highlights previous work in sentiment analysis, emphasizing methods for categorizing sentiment polarity and feature selection. The proposed system introduces a comprehensive approach to sentiment analysis tailored to WhatsApp chats, considering the unique characteristics of conversational data. Key features include user-centric inclusion, nuanced extraction of sentiment, context-aware analysis, and effective handling of conversational nuances. The method section details the proposed system's emphasis on real-time sentiment tracking, user-specific sentiment scores, and integration with WhatsApp's API for broader applications. The results section summarizes the positive outcomes of the proposed method, including user-centric inclusion, nuanced extraction, context-aware illustration, and effective content handling. The future scope section discusses potential enhancements such as real-time sentiment alerts, multilingual support, detailed sentiment analysis, improved user feedback systems, personalized sentiment metrics, and collaboration capabilities. The conclusion highlights the significance of the "WeCare" sentiment analysis project in improving our understanding of digital conversations, particularly within the context of WhatsApp chats. The project's commitment to user-centric design, adaptability, and ethical considerations positions it as a dynamic tool for enhancing digital interactions and well-being. Finally, acknowledgments recognize the contributions of the development team and academic mentor.

Nowcasting Euro area GDP with news sentiment: A tale of two crises

SummaryThis paper shows that newspaper articles contain signals that can materially improve real‐time nowcasts of real GDP growth for the Euro area. Using articles from 15 popular European newspapers, which are machine translated into English, we create sentiment metrics that update daily and assess their value for nowcasting, comparing with competitive and rigorous benchmarks. We find that newspaper text is especially helpful early in the quarter before other indicators are available. We also find that general‐purpose sentiment measures perform better than more economics‐focused ones in response to unanticipated events and nonlinear supervised models can help capture extreme movements in growth but require sufficient training data to be effective.

Lost in translation. When sentiment metrics for one market are derived from two different languages

We compare two text-based proxies for the sentiment of investors in the Japanese market. Both proxies are constructed by Thomson Reuters using the same algorithm, and the only difference between them is that the first proxy is derived using only Japanese language items; the second is derived from English language items. The correlation between the proxies is low and this suggests that they measure different aggregate affective states. The English-language sentiment proxy is found to have a positive and statistically significant association with Japanese returns before a key date in the history of Abenomics relating to quantitative easing (April 2013). After this date, only the Japanese-language proxy has a positive and statistically significant association with the market. Studies of sentiment are predominantly based on proxies derived from English language sources; our results contribute to this nascent research field and suggest that not using both the English and local-language sources when constructing text-based sentiment indices may lead to misleading or mistaken inferences.

Sentiment and Presentiment in Twitter: Do Trends in Collective Mood “Feel the Future”?

Meta-analyses of experiments investigating human behavioral and physiological reactions to unpredictable future events suggest the existence of a poorly understood ability to “feel the future.” Is this effect reflected in sentiment metrics based on social media posts? To find out, analysis of 13 years of daily Twitter sentiment data in 10 languages was examined two weeks prior to events assessed as significantly negative and unpredictable, including acts of terrorism, mass shootings, unexpected deaths of celebrities, etc. Results of the analysis was statistically significant (p = 0.001), suggesting the existence of a form of collective presentiment.

Investor sentiment metrics and stocks market returns: a study of the causality relationship using VAR models

Investor sentiment metrics and stocks market returns: a study of the causality relationship using VAR models

Investor sentiment metrics and stock market returns: a study of the causality relationship using VAR models

Investor sentiment metrics and stock market returns: a study of the causality relationship using VAR models

Analysis of Social Media Impact on Stock Price Movements Using Machine Learning Anomaly Detection

The massive increase in the volume of data generated by individuals on social media microblog platforms such as Twitter and Reddit every day offers researchers unique opportunities to analyze financial markets from new perspectives. The meme stock mania of 2021 brought together stock traders and investors that were also active on social media. This mania was in good part driven by retail investors’ discussions on investment strategies that occurred on social media platforms such as Reddit during the COVID-19 lockdowns. The stock trades by these retail investors were then executed using services like Robinhood. In this paper, machine learning models are used to try and predict the stock price movements of two meme stocks: GameStop ($GME) and AMC Entertainment ($AMC). Two sentiment metrics of the daily social media discussions about these stocks on Reddit are generated and used together with 85 other fundamental and technical indicators as the feature set for the machine learning models. It is demonstrated that through the use of a carefully chosen mix of a meme stock’s fundamental indicators, technical indicators, and social media sentiment scores, it is possible to predict the stocks’ next-day closing prices. Also, using an anomaly detection model, and the daily Reddit discussions about a meme stock, it was possible to identify potential market manipulators.

TRIVIA’s social media effectiveness

Learning outcomes Besides the metrics developed to measure the impact of the video campaign run by the company, the case has serious practical implications for all companies in emerging markets selling budget/low-cost products. Marketing managers can take note of these key performance indicators in evaluating the success of their social media campaigns.More importantly, students need to understand how the campaign managed to integrate uninitiated and less-privileged consumers to the world of social media with aspirational sentiments. Case overview/synopsis The case deals with a digital media planning activity for a low-cost handset company based in India. TRIVIA International Ltd is a manufacturer and marketer of smart phones and feature phone in the price range of below US$70–100 and US$20, respectively. The phones cater to low-income consumer bracket, which forms the vast microcosm of India. The consumer base is huge, but the purchasing power is very low, so they are at times referred to as the micro-consumer in Bottom of Pyramid approach. To approach this consumer base, Trivia planned a very engaging campaign on social media that yielded positive results, contrary to popular belief that only high-income individuals explore the social media intensively through their smart phones. The chapter ends with a set of recommendations for all digital managers who want to analyse their campaigns effectively via awareness, sentiment and engagement metrics. Complexity academic level The case is meant for all post graduate programs in Business Management, which include the MBAs, Masters in Business Administration and the Executive Development program for managers. The case can also be used in Part Time Business Management programs held over weekends for working professionals. Most likely the students are going to handle this case in their second year of a full-time program or at an advanced level of their career development programs. The subjects so mentioned here were offered as electives or areas of specialization. It is presumed that the students undergoing these courses have basic understanding of marketing management. Supplementary materials Teaching notes are available for educators only. Subject code CSS 8: Marketing.

News Sentiment and the Risk of a Stock Price Crash Risk: Based on Financial Dictionary Combined BERT‐DCA

This study combines a financial knowledge dictionary and pretraining method based on BERT (Bidirectional Encoder Representation from Transformers) to construct a deep learning model for identifying stock news sentiments. The study then calculates the sentiment metrics of all stocks and analyzes the impact of news sentiment on the risk of a stock price crash and its heterogeneity. The results show that stocks with more positive sentiment metrics have a higher risk of crash in the following year. We also investigate the information intermediation and investor sentiment channels by which news sentiment affects the risk of a crash. The results show that more net insider sales, lower information transparency, and less analyst coverage amplify the impact of news sentiment on future crash risk, which is consistent with the information intermediation channel. Additionally, more retail investor positions, more active investor sentiment, and divergence between analysts’ opinions and news amplify the impact of news sentiment on the risk of a future stock price crash, which is consistent with the investor sentiment channel.

Сентимент частных инвесторов в объяснении различий в биржевых характеристиках акций российского рынка

In our paper, for the first time, we examine the influence of the sentiment of private investors in social networks on the trade characteristics of stocks in the Russian market. Monthly return rates and trading volumes are analyzed under the control of financial indicators and indicators of the quality of corporate governance of stock issuers, as well as the changing external environment in the period from 2013 to 2020. The sample for various sentiment metrics is based on unique data: messages in the Telegram and mfd.ru platforms. The tonality of messages is diagnosed according to the authors’ method using artificial intelligence (neural network). The main conclusion is: the sentiment can be seen as an explanatory factor in pricing and trading activity. The influence of sentiment is non-linear. The author’s HYPE indicator of sentiment is proposed and compared in terms of explanatory ability of the trade characteristics with a wide range of proxy variables. The explanatory ability to identify differences is realized through regression constructions on panel data. It is shown that trade characteristics are more sensitive to the growth of negative messages, which is consistent with the postulates of behavioral finance. An increase in messages’ number of both positive and negative sentiment contributes to the growth of trading activity. An important practical conclusion is: following the crowd when the company is most intensely discussed will not result in high returns to an investor.

Investor sentiment in the equity market and investments in corporate-bond funds

This study explores the relation between investor sentiment in equity market and investments in corporate-bond funds. Investors tend to move into and out of corporate-bond funds when contemporaneous sentiment in equity market differs from the historical average. Specifically, a one-standard-deviation decrease in equity-market sentiment generates 0.1% and 0.4% inflows for active and index funds, respectively. It reflects the time-varying flight-to-safety behavior of investors. Besides, funds with low exposure to equity-market sentiment appear to attract inflows and funds with high exposure to equity-market sentiment experience outflows, indicating that investors are likely to avoid sentiment risk. Morever, the result shows that funds with the highest negative sentiment exposure significantly outperform the funds with the highest positive sentiment exposure by 2.22%–2.52% per annum. The results are robust to using alternative sentiment metrics and considering different subperiods.

Specialists, Scientists, and Sentiments: Word2Vec and Doc2Vec in Analysis of Scientific and Medical Texts.

Analyze performance of unsupervised embedding algorithms in sentiment analysis of knowledge-rich data sets. We apply state-of-the-art embedding algorithms Word2Vec and Doc2Vec as the learning techniques. The algorithms build word and document embeddings in an unsupervised manner. To assess the algorithms’ performance, we define sentiment metrics and use a semantic lexicon SentiWordNet (SWN) to establish the benchmark measures. Our empirical results are obtained on the Obesity data set from i2b2 clinical discharge summaries and the Reuters Science dataset. We use the Welch’s test to analyze the obtained sentiment evaluation. On the Obesity data, the Welch’s test found significant difference between the SWN evaluation of the most positive and most negative texts. On the same data, the Word2Vec results support the SWN results, whereas the Doc2Vec results partially correspond to the Word2Vec and the SWN results. On the Reuters data, the Welch’s test did not find significant difference between the SWN evaluation of the most positive and most negative texts. On the same data, Word2Vec and Doc2Vec results only in part correspond to the SWN results. In unsupervised sentiment analysis of medical and scientific texts, the Word2Vec sentiment analysis has been more consistent with the SentiWordNet sentiment assessment than the Doc2Vec sentiment analysis. The Welch’s test of the SentiWordNet results has been a strong indicator of future correspondence between Word2Vec and SentiWordNet results.

Firm-specific investor sentiment for the Chinese stock market

Investor sentiment plays a significant role in asset prices, and sieves out safer and speculative stocks, and captures institutional and retail investors’ demand shocks. Previous studies focus on the effects of investor sentiment in the developed stock markets using various sentiment metrics. We propose overnight and over-weekend returns as firm-specific investor sentiment (FSIS) to reflect retail investors’ beliefs about the developing Chinese stock market, and show that FSIS has short-term persistence, a negative relation with intraday returns and price impact, a U-shaped relation with trading activity, an inverse U-shaped relation with long-run performance and a positive impact on cross-sectional returns. FSIS captures the characteristics of the Chinese stock market with short-selling constraints, and market-level information of FSIS explains risk premiums from sentiment-driven mispricing. Our results also reflect the stronger-self remediability after the 2015 crash of the Chinese stock market.

Forecasting Stock Price Changes Using Natural Language Processing

All investors attempt to predict stock market returns when they make investment decisions. However, making such predictions is not a trivial task. As a result, many strategies have been proposed by researchers as potential ways to predict stock returns. More recently, data analytics - in general - and natural language processing, in particular, have been identified as viable options. In this project, we investigate the use of natural language processing to forecast stock price changes. Specifically, we analyze firms’ 10-K and 10-Q reports to identify sentiment. Using the computed sentiment scores, we develop models to predict the direction of stock price movements both in the short run and in the long run. Our first step in developing these models is to investigate some sentiment scoring methods and apply them to the Loughran-McDonald dictionary. Next, we use the model word2vec to extend the usage of the Loughran-McDonald dictionary and then apply the sentiment metrics. Additionally, we apply the proposed sentiment metrics to FinBERT, which learns contextual relations between words. Finally, we build supervised machine learning algorithms that use the proposed sentiments as inputs to forecast price changes. We train our algorithms on 10-K and 10-Q reports of 48 companies in the S&P 500 from 2013 to 2017. Finally, we test our models on the corresponding reports from 2018 to 2019 and conclude that predictive signals can be extracted from 10-K and 10-Q reports.

Nowcasting Euro Area GDP with News Sentiment: A Tale of Two Crises

This paper shows that newspaper articles contain timely economic signals that can materially improve nowcasts of real GDP growth for the euro area. Our text data is drawn from fifteen popular European newspapers, that collectively represent the four largest Euro area economies, and are machine translated into English. Daily sentiment metrics are created from these news articles and we assess their value for nowcasting. By comparing to competitive and rigorous benchmarks, we find that newspaper text is helpful in nowcasting GDP growth especially in the first half of the quarter when other lower-frequency soft indicators are not available. The choice of the sentiment measure matters when tracking economic shocks such as the Great Recession and the Great Lockdown. Non-linear machine learning models can help capture extreme movements in growth, but require sufficient training data in order to be effective so become more useful later in our sample. JEL Classification: C43, C45, C55, C82, E37

#Dementia: An Evaluation of the Worldwide Dementia Twitter Network

Social media engages an international network of healthcare stakeholders. Our study characterized the #dementia community on Twitter, which we hypothesized has increased substantially in the last six years. Symplur Signals, a healthcare social media analytics platform, was used to analyze public #dementia tweets between 1/1/2014 and 12/31/2019. Spam or users unclassified by the platform were excluded. Tweet activity, content, user characteristics, engagement, associated hashtags, sentiment, and network analysis metrics were obtained. Exactly 2,149,494 tweets made by 105,938 users resulted in 21,460,413,770 impressions. There was a 44.6% and 29.1% compound annual growth rate of tweets and users, respectively. The five most frequently tweeted words were people, care, help, living, and support. Among users with identifiable locations, the United Kingdom, United States, and Canada had the greatest number of users. Healthcare associated individuals (17.8%), health advocacy organizations (17.6%), non-healthcare associated individuals (17.0%), and non-healthcare associated organizations (13.8%) were among the top 500 influencers based on tweet number. Hashtags related to caregiving (#Caregiving, #carers, #care, #caregivers, #caregiver, #dementiacare) and Alzheimer’s disease (#alzheimers, #Alzheimer, #Alz, #EndAlz) were among the top 15 associated hashtags. Analysis of the 1,000 most recent tweets shows more tweets with positive (71.0%) than negative (29.0%) sentiments. Network analysis mapping connections between users shows advocacy organizations (i.e. @ASAging) and other health individuals (i.e. @LEAD_Coalition) were central conversation hubs in 2019. The #dementia network discusses themes important to older adults. The network’s rapid growth has enabled increased dissemination of dementia-related information by both healthcare and non-healthcare associated individuals and organizations.

Buzzwords build momentum: Global financial Twitter sentiment and the aggregate stock market

We examine the long-term relationship between signals derived from nine years of unstructured social media microblog text data and financial market developments in five major economic regions. Employing statistical language modeling techniques we construct directional sentiment metrics and link these to aggregate stock index returns. To address the noise in finance-related Twitter messages we identify expert users whose tweets predominantly focus on finance topics. We document that expert users are the main drivers behind an interdependence between Twitter sentiment and financial markets. The direct prediction value of expert sentiment metrics for stock index returns, however, is found to be elusive and short-lived. Yet, we detect significant predictive gains over benchmark models in times of negative market returns. In consequence, the relation between expert sentiment metrics and stock indices is sufficient to devise hypothetically profitable cross-sectional as well as time series momentum investment strategies for futures based on Twitter signals that survive basic transaction cost assumptions. In this context, our results show that expert sentiment signals can yield higher risk-adjusted returns than classical price-based signals.

Sentiment Metrics and Investor Demand

ABSTRACTRecent work suggests that sentiment traders shift from safer to more speculative stocks when sentiment increases. Exploiting these cross‐sectional patterns and changes in share ownership, we find that sentiment metrics capture institutional rather than individual investors’ demand shocks. We investigate the underlying economic mechanisms and find that common institutional investment styles (e.g., risk management, momentum trading) explain a significant portion of the relation between institutions and sentiment.

Buzzwords Build Momentum: Global Financial Twitter Sentiment and the Aggregate Stock Market

We examine the long-term relationship between signals derived from nine years of unstructured social media microblog text data and financial market developments in five major economic regions. Employing statistical language modeling techniques we construct directional sentiment metrics and link these to aggregate stock index returns. To address the noise in finance-related Twitter messages we identify expert users whose tweets predominantly focus on finance topics. We document that expert users are the main drivers behind a significant contemporaneous link between Twitter sentiment and financial markets. Notably, the link remains equally strong in times of major news events impacting the markets. The direct prediction value of expert sentiment metrics for stock index returns, however, is found to be elusive and short-lived. Yet, the relation between expert sentiment metrics and stock indices is sufficient to devise profitable cross-sectional as well as time series momentum investment strategies based on Twitter signals that survive basic transaction cost assumptions. In this context, our results show that expert sentiment signals can yield higher risk-adjusted returns than classical price-signals.