Essays on momentum investing strategies

The momentum investment strategy, which buys recent winner stocks and sells recent loser stocks, earns returns that are simply too good to be explained by traditional finance theories. This thesis extends our understanding of the sources of momentum profits. The research shows that part of the seemingly anomalous returns can be explained by the market's reaction to public news, is affected by how delisting returns are calculated, and is biased by ignoring the time-varying risk of the trading strategy.

Information flow, trading activity and commodity futures volatility

Based on unique news data relating to gold and crude oil, we investigate how news volume and sentiment, shocks in trading activity, market depth and trader positions unrelated to information flow covary with realized volatility. Positive shocks to the rate of news arrival, and negative shocks to news sentiment exhibit the largest effects. After controlling for the level of news flow and cross-correlations, net trader positions play only a minor role. These findings are at odds with those of [Wang (2002a). The Journal of Futures Markets, 22, 427–450; Wang (2002b). The Financial Review, 37, 295–316], but are consistent with the previous literature which doesn't find a strong link between volatility and trader positions.

A pilot study on affective classification of facial images for emerging news topics

The proliferation of news reports published in online websites and news information sharing among social media users necessitates effective techniques for analysing the image, text and video data related to news topics. This paper presents the first study to classify affective facial images on emerging news topics. The proposed system dynamically monitors and selects the current hot (of great interest) news topics with strong affective interestingness using textual keywords in news articles and social media discussions. Images from the selected hot topics are extracted and classified into three categorized emotions, positive, neutral and negative, based on facial expressions of subjects in the images. Performance evaluations on two facial image datasets collected from real-world resources demonstrate the applicability and effectiveness of the proposed system in affective classification of facial images in news reports. Facial expression shows high consistency with the affective textual content in news reports for positive emotion, while only low correlation has been observed for neutral and negative. The system can be directly used for applications, such as assisting editors in choosing photos with a proper affective semantic for a certain topic during news report preparation.

The direction of media influence: Real-estate news and the stock market

This paper uses a novel identification strategy to test the influence of news media on the stock market. Because the stock market does not impact the media coverage of the housing market, a relationship between real-estate news and shares of companies engaged in the housing market is attributable media influence. I find that the content of reporting exhibits a significant relationship with stock returns, and the amount of news with the number of trades. These relationships exist even after controlling for known risk factors, housing market performance and intra-week correlation. This finding is consistent with the function of the media as a source of information and sentiment in financial markets.

Are investors guided by the news disclosed by companies or by journalists?

Most previous studies demonstrating the influential role of the textual information released by the media on stock market performance have concentrated on earnings-related disclosures. By contrast, this paper focuses on disposal announcements, so that the impacts of listed companies’ announcements and journalists’ stories can be compared concerning the same events. Consistent with previous findings, negative words, rather than those expressing other types of sentiment, statistically significantly affect adjusted returns and detrended trading volumes. However, extending previous studies, the results of this paper indicate that shareholders’ decisions are mainly guided by the negative sentiment in listed companies’ announcements rather than that in journalists’ stories. Furthermore, this effect is restricted to the announcement day. The average market reaction–measured by adjusted returns–is inversely related only when the announcements are ignored by the media, but the dispersion of market reaction–measured by detrended trading volume–is positively affected only when announcements are followed up by journalists.

Event extraction using behaviors of sentiment signals and burst structure in social media

Significant world events often cause the behavioral convergence of the expression of shared sentiment. This paper examines the use of the blogosphere as a framework to study user psychological behaviors, using their sentiment responses as a form of ‘sensor’ to infer real-world events of importance automatically. We formulate a novel temporal sentiment index function using quantitative measure of the valence value of bearing words in blog posts in which the set of affective bearing words is inspired from psychological research in emotion structure. The annual local minimum and maximum of the proposed sentiment signal function are utilized to extract significant events of the year and corresponding blog posts are further analyzed using topic modeling tools to understand their content. The paper then examines the correlation of topics discovered in relation to world news events reported by the mainstream news service provider, Cable News Network, and by using the Google search engine. Next, aiming at understanding sentiment at a finer granularity over time, we propose a stochastic burst detection model, extended from the work of Kleinberg, to work incrementally with stream data. The proposed model is then used to extract sentimental bursts occurring within a specific mood label (for example, a burst of observing ‘shocked’). The blog posts at those time indices are analyzed to extract topics, and these are compared to real-world news events. Our comprehensive set of experiments conducted on a large-scale set of 12 million posts from Livejournal shows that the proposed sentiment index function coincides well with significant world events while bursts in sentiment allow us to locate finer-grain external world events.

The good news about television: attitudes aren't getting worse: tracking public attitudes toward TV advertising

Periodically tracking public sentiment toward television advertising (TVA) is an important barometer for the advertising industry and its myriad stakeholders. To date, however, most studies of consumers’ attitudes to TVA have been cross-sectional. This study, alternatively, provides a quasi-longitudinal examination of Australian attitudes toward TVA across four time points (2002, 2005, 2008, and 2010). Findings suggest that although attitudes toward TVA are generally negative, in fact they have not deteriorated over time. Considerable scope consequently exists for improving consumer attitudes toward TVA.

Text analytics of social media: Sentiment analysis, event detection and summarization

In the last decade, large numbers of social media services have emerged and been widely used in people's daily life as important information sharing and acquisition tools. With a substantial amount of user-contributed text data on social media, it becomes a necessity to develop methods and tools for text analysis for this emerging data, in order to better utilize it to deliver meaningful information to users. ^ Previous work on text analytics in last several decades is mainly focused on traditional types of text like emails, news and academic literatures, and several critical issues to text data on social media have not been well explored: 1) how to detect sentiment from text on social media; 2) how to make use of social media's real-time nature; 3) how to address information overload for flexible information needs. ^ In this dissertation, we focus on these three problems. First, to detect sentiment of text on social media, we propose a non-negative matrix tri-factorization (tri-NMF) based dual active supervision method to minimize human labeling efforts for the new type of data. Second, to make use of social media's real-time nature, we propose approaches to detect events from text streams on social media. Third, to address information overload for flexible information needs, we propose two summarization framework, dominating set based summarization framework and learning-to-rank based summarization framework. The dominating set based summarization framework can be applied for different types of summarization problems, while the learning-to-rank based summarization framework helps utilize the existing training data to guild the new summarization tasks. In addition, we integrate these techneques in an application study of event summarization for sports games as an example of how to better utilize social media data. ^

Text Analytics of Social Media: Sentiment Analysis, Event Detection and Summarization

In the last decade, large numbers of social media services have emerged and been widely used in people's daily life as important information sharing and acquisition tools. With a substantial amount of user-contributed text data on social media, it becomes a necessity to develop methods and tools for text analysis for this emerging data, in order to better utilize it to deliver meaningful information to users. Previous work on text analytics in last several decades is mainly focused on traditional types of text like emails, news and academic literatures, and several critical issues to text data on social media have not been well explored: 1) how to detect sentiment from text on social media; 2) how to make use of social media's real-time nature; 3) how to address information overload for flexible information needs. In this dissertation, we focus on these three problems. First, to detect sentiment of text on social media, we propose a non-negative matrix tri-factorization (tri-NMF) based dual active supervision method to minimize human labeling efforts for the new type of data. Second, to make use of social media's real-time nature, we propose approaches to detect events from text streams on social media. Third, to address information overload for flexible information needs, we propose two summarization framework, dominating set based summarization framework and learning-to-rank based summarization framework. The dominating set based summarization framework can be applied for different types of summarization problems, while the learning-to-rank based summarization framework helps utilize the existing training data to guild the new summarization tasks. In addition, we integrate these techneques in an application study of event summarization for sports games as an example of how to better utilize social media data.

How does ‘Newstainment’ actually work? : ethnographic research methods and contemporary popular news

Much debate has taken place recently over the potential for entertainment genres and unorthodox forms of news to provide legitimate – indeed democratized – in-roads into the public sphere. Amidst these discussions, however, little thought has been paid to the audiences for programs of this sort, and (even when viewers are considered) the research can too easily treat audiences in homogenous terms and therefore replicate the very dichotomies these television shows directly challenge. This paper is a critical reflection on an audience study into the Australian morning “newstainment” program Sunrise. After examining the show and exploring how it is ‘used’ as a news source, this paper will promote the use of ethnographic study to better conceptualize how citizens integrate and connect the increasingly fragmented and multifarious forms of postmodern political communication available in their everyday lives.

User behaviours and intentions with digital news media in Australia

This paper summarises findings from a survey of user behaviors and intentions towards digital media and information in Australia. It was undertaken in the first quarter of 2009 by the Queensland University of Technology Creative Industries Faculty and was funded by the Smart Services Cooperative Research Centre. The survey targeted users of 2 news and information sites that are available online only. Findings highlighted differences between the 18-24 year age segment and older users. Social networks (specifically friends and family) were rated as the least reliable, relevant and accurate sources of news. Other findings indicate online news sources that are associated with an established newspaper are highly valued as reliable, relevant and accurate news sources by most people. While most people prefer to use online news sources, there is a great deal of variation in the ways in which people actually use online news. From a total of 524 respondents to the survey it was possible to identify three main types of online news consumers: convenience, loyal and customising users.

Comparing international television news : the case of violence presentation in Germany, Great Britain and Russia -- [in German : Gewalt in internationalen Fernsehnachrichten - eine komparative Analyse medialer Gewaltpräsentation in Deutschland, Großbritannien und Russland]

This is an important book that ought to launch a debate about how we research our understanding of the world, it is an innovative intervention in a vital public issue, and it is an elegant and scholarly hard look at what is actually happening. Jean Seaton, Prof of Media History, U of Westminster, UK & Official Historian of the BBC -- Summary: This book investigates the question of how comparative studies of international TV news (here: on violence presentation) can best be conceptualized in a way that allows for crossnational, comparative conclusions on an empirically validated basis. This book shows that such a conceptualization is necessary in order to overcome existing restrictions in the comparability of international analysis on violence presentation. Investigated examples include the most watched news bulletins in Great Britain (10o'clock news on the BBC), Germany (Tagesschau on ARD) and Russia (Vremja on Channel 1). This book highlights a substantial cross-national violence news flow as well as a cross-national visual violence flow (key visuals) as distinct transnational components. In addition, event-related textual analysis reveals how the historical rootedness of nations and its symbols of power are still manifested in televisual mediations of violence. In conclusion, this study lobbies for a conscientious use of comparative data/analysis both in journalism research and practice in order to understand what it may convey in the different arenas of today’s newsmaking.

Audience & Market Foresight : Consumer Trends in Digital News and Information in Australia

We all know that the future of news is digital. But mainstream news providers are still grappling with how to entice more customers to digital news. This paper provides context for a survey currently underway on user intentions towards digital news and entertainment, by exploring: 1. Consumer behaviours and intentions towards digital news and information use; 2. Current trends in the Australian online news and information sector; 3. Issues and emerging opportunities in the Australian (and global) environment. Key influences on digital use of news and information are pricing and access. The paper highlights emerging technical opportunities and flags service gaps as at December 2008. These gaps include multiple disconnects between: 1. Changing user intentions towards online and location based news (news based on a specific locality as chosen by the user) and information; 2. The ability by consumers to act on these intentions via the availability and cost of technologies; 3. Younger users prefer entertainment to news; 4. Current digital offerings of traditional news providers and opportunities. These disconnects present an opportunity for online news suppliers to appraise and resolve. Doing so may enhance their online news and information offering, attract consumers and improve loyalty. Outcomes from this paper will be used to identify knowledge gaps and contribute to the development of further analysis on Australian consumers and their behaviours and intentions towards online news and information. This will be ndertaken via focus groups as part of a broader study by researchers at the Creative Industries Faculty at the Queensland University of Technology supported by the Smart Services Cooperative Research Centre.