Personal tools

Information zum Seitenaufbau und Sprungmarken fuer Screenreader-Benutzer: Ganz oben links auf jeder Seite befindet sich das Logo der JLU, verlinkt mit der Startseite. Neben dem Logo kann sich rechts daneben das Bannerbild anschließen. Rechts daneben kann sich ein weiteres Bild/Schriftzug befinden. Es folgt die Suche. Unterhalb dieser oberen Leiste schliesst sich die Hauptnavigation an. Unterhalb der Hauptnavigation befindet sich der Inhaltsbereich. Die Feinnavigation findet sich - sofern vorhanden - in der linken Spalte. In der rechten Spalte finden Sie ueblicherweise Kontaktdaten. Als Abschluss der Seite findet sich die Brotkrumennavigation und im Fussbereich Links zu Barrierefreiheit, Impressum, Hilfe und das Login fuer Redakteure. Barrierefreiheit JLU - Logo, Link zur Startseite der JLU-Gießen Direkt zur Navigation vertikale linke Navigationsleiste vor Sie sind hier Direkt zum Inhalt vor rechter Kolumne mit zusaetzlichen Informationen vor Suche vor Fußbereich mit Impressum

Document Actions

BWL XI: Paper in Expert Systems with Applications

A new research paper has been accepted for publication in Expert Systems with Applications (IF: 4.292). The paper uses machine learning and methods from natural language processing to predict sentence-level polarity labels in financial news.

Title: Predicting Sentence-Level Polarity Labels of Financial News Using Abnormal Stock Returns

Authors: Bernhard Lutz (University of Freiburg), Nicolas Pröllochs, Dirk Neumann (University of Freiburg)

Abstract:

Expert systems for automatic processing of financial news commonly operate at the document level by counting positive and negative term-frequencies. This, however, limits their usefulness for investors and financial practitioners seeking specific positive and negative information on a more fine-grained level. For this purpose, this paper develops a novel machine learning approach for the prediction of sentence-level polarity labels in financial news. The method uses distributed text representations in combination with multi-instance learning to transfer information from the document level to the sentence level. This has two key advantages: (1) it captures semantic information of the textual data and thereby prevents the loss of information caused by bag-of-words approaches; (2) it is solely trained based on historic stock market reactions following the publication of news items without the need for any kind of manual labeling. Our experiments on a manually-labeled dataset of sentences from financial news yield a predictive accuracy of up to 71.20%, exceeding the performance of alternative approaches significantly by at least 5.10 percentage points. Hence, the proposed approach provides accurate decision support for investors and may assist investor relations departments in communicating their messages as intended. Furthermore, it presents promising avenues for future research aiming at studying communication patterns in financial news

Abstract:

Expert systems for automatic processing of financial news commonly operate at the document level by counting positive and negative term-frequencies. This, however, limits their usefulness for investors and financial practitioners seeking specific positive and negative information on a more fine-grained level. For this purpose, this paper develops a novel machine learning approach for the prediction of sentence-level polarity labels in financial news. The method uses distributed text representations in combination with multi-instance learning to transfer information from the document level to the sentence level. This has two key advantages: (1) it captures semantic information of the textual data and thereby prevents the loss of information caused by bag-of-words approaches; (2) it is solely trained based on historic stock market reactions following the publication of news items without the need for any kind of manual labeling. Our experiments on a manually-labeled dataset of sentences from financial news yield a predictive accuracy of up to 71.20%, exceeding the performance of alternative approaches significantly by at least 5.10 percentage points. Hence, the proposed approach provides accurate decision support for investors and may assist investor relations departments in communicating their messages as intended. Furthermore, it presents promising avenues for future research aiming at studying communication patterns in financial news