Sentiment analysis has emerged as a field that has attracted a significant amount of attention since it has a wide variety of applications that could benefit from its results, such as news analytics, marketing, question answering, knowledge management and so on. This area, however, is still early in its development where urgent improvements are required on many issues, particularly on the performance of sentiment classification. In this thesis, three key challenging issues affecting sentiment classification are outlined and innovative ways of addressing these issues are presented. First, text pre-processing has been found crucial on the sentiment classification performance. Consequently, a combination of several existing preprocessing methods is proposed for the sentiment classification process. Second, text properties of financial news are utilised to build models to predict sentiment. Two different models are proposed, one that uses financial events to predict financial news sentiment, and the other uses a new interesting perspective that considers the opinion reader view, as opposed to the classic approach that examines the opinion holder view. A new method to capture the reader sentiment is suggested. Third, one characteristic of financial news is that it stretches over a number of domains, and it is very challenging to infer sentiment between different domains. Various approaches for cross-domain sentiment analysis have been proposed and critically evaluated.

Estratégias do setor financeiro operacionalizadas junto aos jornalistas para construir e garantir presença predominante no noticiário econômico de jornais impressos no Brasil. / The predominance of financial market\'s sources in Brazil\'s economic news and their strategies to consolidate as the only correct economic thought in the media.

Estratégias do setor financeiro operacionalizadas junto aos jornalistas para construir e garantir presença predominante no noticiário econômico de jornais impressos no Brasil. / The predominance of financial market\'s sources in Brazil\'s economic news and their strategies to consolidate as the only correct economic thought in the media.

本研究針對台灣地區投資人情緒程度與財經新聞之間的關聯進行研究。本研究以台灣地區發布之恐慌指數(TWVIX)做為投資人情緒程度代理變數。研究樣本期間為2006年12月至2014年12月，新聞資料來源為知識贏家(KMW)。本研究使用文字分析法針對財經新聞內容進行分析，透過計算詞頻和建立語調評分方法來衡量新聞內涵。 本研究之主要發現有：單日媒體發布之新聞數量與恐慌指數呈正相關、單篇新聞平均長度與恐慌指數呈負相關。在文字內容分析的迴歸結果方面，實證結果指出當新聞內的正向字詞使用的越多，投資人情緒程度會降低，相反地，當負向字詞使用的越多，將會提升投資人情緒，產生恐慌狀態。另外，結果也發現當新聞的語調越接近樂觀情緒時，投資人的情緒程度會降低。 本文嘗試了一個新的切入角度來討論金融指標，也就是由新聞媒體效果來探討與恐慌指數之關聯性。本研究結果提供了實證證據證明台灣地區媒體可有效地透過控制財經新聞的內涵，來影響市場上投資人的情緒程度。 / This study examines the relation between investor sentiment and financial news released in Taiwan for the period from Dec. 2006 to Dec. 2014. This study adopts the TAIEX Options Volatility Index (TWVIX) as a proxy for investor sentiment. The study employs the text analysis approach to measure the tone of financial news. The news data is collected from Knowledge Management Winner (KMW) of China Times. The empirical results show that the number of news released on a day has a positive effect on TWVIX; otherwise the length of financial news affects TWVIX negatively. By the text analysis, this study finds that with more positive wording, the degree of investor sentiment is decreased. On the contrary, the quantity of negative words provokes the investor sentiment. Lastly, this research also finds that when the tone of financial news is closer to optimism, it help ease the level of investor sentiment. This research explores a new cut point to discuss the VIX, connecting the issue between a principal financial index and news media in Taiwan. This study believes that the results provide the evidence that Taiwan media actually communicate messages to investors effectively. Thus, news media has the ability to influence investor sentiment no matter by news coverage, words usage and the tone of financial news.

Financial data in banks are unstructured and complicated. It is challenging to analyze these texts manually due to the small amount of labeled training data in financial text. Moreover, the financial text consists of language in the economic domain where a general-purpose model is not efficient. In this thesis, data had collected from MFN (Modular Finance) financial news, this data is scraped and persisted in the database and price indices are collected from Bloomberg terminal. Comprehensive study and tests are conducted to find the state-of-art results for classifying the sentiments using traditional classifiers like Naive Bayes and transfer learning models like BERT and FinBERT. FinBERT outperform the Naive Bayes and BERT classifier. The time-series indices for sentiments are built, and their correlations with price indices calculated using Pearson correlation. Augmented Dickey-Fuller (ADF) is used to check if both the time series data are stationary. Finally, the statistical hypothesis Granger causality test determines if the sentiment time series helps predict price. This result shows that there is a significant correlation and causal relation between sentiments and price.

No description available.

In late 2019, the spread of the coronavirus SARS-CoV-2 (Severe acute respiratory syndrome coronavirus 2) began. The viral disease, also known as covid-19, started spreading from China to large parts of the world in early 2020, resulting in a large number of cases, deaths, as well as major impacts on the economy of nations, organizations, and individuals alike. In order to limit the spread of the virus, several pharmaceutical companies, including Pfizer and Moderna, initiated a vaccine development. This process led pharmaceutical companies to communicat ea large amount of corporate news to investors, among other stakeholders. Previous research has shown how the stock market responds to corporate news. However, prior to this study, a lack of research on how the stock market behaves in relation to corporate news announced by companies under difficult global conditions was identified. The aim of this study is to investigate the stock market's reaction to press releases by Moderna and Pfizer regarding their vaccine development for covid-19. The methods used to study the subject are based on analysis of historical share price data and the calculation of abnormal returns. The results have to some extent indicated that selected press releases have played a significant role as an influencer in relation to investor sentiment and the stock market’s behavior. In terms of the result as a whole, the number of significant days is not sufficient to support the alternative hypothesis. Market reactions in the share prices were identified on the dates surrounding the press releases regarding the vaccine-development, but since only 22 percent of the abnormal yield was significant, the result did not appear to be aligned with the alternative hypothesis. / Under slutet av år 2019 började spridningen av coronaviruset SARS-CoV-2 (Severe acute respiratory syndrome coronavirus 2). Virussjukdomen, i folkmun känd som covid-19, började under år 2020 spridas från Kina till stora delar av övriga världen, vilket har resulterat i ett stort antal insjuknade, dödsfall och även en påverkan på såväl nationers som organisationers och individers ekonomi. För att begränsa spridningen av viruset påbörjade ett flertal läkemedelsbolag en vaccinframtagning. Denna process ledde till att läkemedelsbolagen kommunicerade en stor mängd bolagsnyheter till omvärlden, däribland investerare. Tidigare forskning har visat hur aktiemarknaden svarat på bolagsnyheter, men inför denna studie identifierades en brist på forskning kring hur aktiemarknaden agerar i förhållande till bolagsnyheter under extrema världssituationer likt coronaviruspandemin. Syftet med denna studie är att undersöka marknadens reaktion till följd av pressmeddelanden från företagen Moderna och Pfizer rörande vaccinutvecklingen för covid-19. Reaktionen studeras med hjälp av historiska aktiekursdata och beräkning av onormal avkastning. Resultaten har till en viss del indikerat på att utvalda pressmeddelanden har spelat en signifikant roll som påverkande faktor gentemot aktiemarknaden. Sett till resultatet i sin helhet är antalet signifikanta dagar inte tillräckligt för att utgöra stöd till alternativhypotesen. Marknadsreaktioner kunde identifieras till följd av pressmeddelanden om vaccinutvecklingen, men eftersom enbart 22 procent av den onormala avkastningen var signifikant så föreföll inte resultatet i linje med alternativhypotesen.

新聞報導在現今社會中為最便利與即時的資訊來源，也是公司與管理階層用來發布年度表現與其他重要消息的管道。本研究想要去找出新聞報導的內容與股票市場的反應之間的關連性，以及財務報表附註提供比新聞報導更多的其他資訊。 本研究以2013年度臺灣90家上市公司為樣本去探討股票市場與新聞報導及財報附註之間的關係。藉由內容分析法的方式去評估新聞報導與財報附註中的正向、負向情緒以及兩者之間的資訊差異以及兩者對股票市場的影響。藉由兩種不同的資訊來源，找出股票市場對於此兩者是否有不同的反應，並從中判斷兩種資訊內容的差異。 本研究發現新聞報導中的正向情緒與股票市場有正向的關連性，然而無法顯著證明在市場反應方面，財報附註能提供比新聞報導更豐富的資訊。 / Financial news articles are the most convenient and timely information in today’s world. Companies and managers can announce current year performance and other concurrent disclosures to investors and stakeholders by financial news articles. This study hopes to find the relationships between financial news articles and the stock market and the information content gap between financial news articles and footnotes to financial statements. This study uses 90 listed companies in TSE of 2013 to test stock market responses to the information content in the financial news articles and footnotes to financial statements. This study implements content-analysis technique, which count words for characterizing as optimistic and negative tone and helps to decide the abundancy of information content elaborated, to find the information content to stock market reactions and compare the information content gap between financial news articles and footnotes to financial statements. This study finds that optimistic sentiment expressed in financial news articles positively relates to the stock price movement. However, this study is unable to reach a conclusion that information disclosed in the footnotes to financial statements is significantly enough to represent the existence of information content gap as compared to financial news articles.