Business Intelligence 2.0 - Knowledge-based Services for Analyzing Opinion Formation on Web 2.0

Abstract

Motivation

More and more customers engage in online social communities for the purpose of networking and discussing their opinions with other members. The interactive opinion exchange has a high impact on the purchasing decision of fellow consumers. Analyzing opinion formation on Web 2.0 is of vital importance for companies, as they are then in a position to judge opportunities and risks at an early stage and to initiate marketing measures for influencing opinion formation. The multitude of text-based consumer opinions and the complexity of social interactions require an automated analysis and decision support

This work introduces business intelligence services for supporting market research and marketing. Social opinion formation processes on Web 2.0 are analyzed by mining services and continuously supervised by monitoring services. Early warning services ensure that marketing managers are alerted at an early stage when critical situations arise. Decision support services help selecting appropriate measures for influencing opinion formation.

Contribution

This work’s contribution to research is a static and dynamic analysis of opinion formation on the basis of real data by taking the opinions expressed in postings as well as the social interactions of discussion members into account. Existing work focuses on single aspects of opinion formation, concentrates on analyzing past data, and mainly relies on theoretical models. This work, in contrast, enables case-based decision support for actual situations.

Application

On the basis of the paradigm of design science, new innovative business intelligence services are developed and evaluated with the aid of case studies. The business intelligence services are applied to analyze opinions towards three different kinds of products: soccer shoes, smart phones and computer games. Choosing physical, hybrid and digital products from different industries demonstrates the broad ranges of applications and the high benefits to market research and marketing.

Description

Overview

This work answers the question of how valuable knowledge on online opinion formation can be gained and monitored in order to enable an early warning in critical situations and a decision support for influencing opinion formation.

Mining

Opinions can be found in a multitude of text-based contributions. First, text mining techniques are employed in order to gain the opinions in an automated manner. With the aid of support vector machines, consumers’ opinions on products and product features are identified. Next, coherences between these opinions are detected by data mining methods. Association rules and decision trees reveal the dependencies among the evaluations of product features and detect the main influencing factors on the overall evaluations of products. In order to gain knowledge on the formation of opinions, the overall discussion thread consisting of a sequence of postings must be considered, thus providing a deeper insight into the social interaction of the discussion members. By taking the overall discussion network into account, opinion leaders and trends can be detected with the aid of social network analysis. Focusing on single network members leads to the question of how a member’s opinion can be explained as a consequence of his network relationships. Three influencing factors, i.e. adopting the leader’s opinion, following the opinion of the neighbors and having confidence in one’s own judgment, are analyzed with the aid of methods coming from swarm intelligence.

Monitoring

As opinions change over time, continuous monitoring is crucial. The development of opinion formation is influenced by marketing activities and social interactions within the Internet. Therefore, this work analyzes the development of opinions induced by marketing activities by applying methods from statistics. The influence of marketing campaigns, external events and opinions on competing products is measured and the effects on the sales volume is determined. Moreover, collective opinion formation is examined with respect to group behavior and opinion leadership by applying methods from swarm intelligence.

Early Warning

In order to recognize critical situations in opinion formation at an early stage and to initiate counteractive measures, an early warning system is needed. Situations are considered as critical, when negative opinions are about to spread and to harm the company’s image and sales volume. A neuro fuzzy system is presented which learns rules for identifying critical situations from past experiences. The broad knowledge acquired by mining and monitoring consumers’ opinions enables the warning system to take all relevant influencing factors into account when judging situations.

Decision Support

If critical situations are detected early on, marketing managers are in a position to influence opinion formation. This work supports marketers’ decisions in the field of stealth marketing and influence marketing. On the one hand, what-if analyzes demonstrate which messages should be posted to discussion threads in order to influence opinion formation. A case study reveals the effects achieved when posting consumer experience as well as links to external sources such as professional reviews. On the other hand, what-if analyzes show which opinion leaders should be addressed in order to trigger cascades of influence. The effects of targeting different kinds of opinion leaders in different types of networks are measured and compared.

Project Team

• Project Management: Dipl.-Wirtsch.Inf. Carolin Kaiser
• Students: J. Brückl, J. Kröckel, N. Spassova, M. Weinmann, T. Göhner, A. Crummenauer, K. Schröder, S. Jachtmann, R. Abraham, T. Orgis, S. Winter, S. Frank, C. Schweiger, S. Schlick, F.Bernhard, A. Schwingenschlögl, A. Gerhards, A. Piazza
• Partners: adidas, Hornbach, Psyma, Sony, Glaxo Smith Kline, Dr. Schwabe

Publications

Articles and Papers

• Kaiser, C., Piazza, A., Kröckel, J., Bodendorf, F. (2011). Are Humans like Ants? Analyzing Collective Opinion Formation in Online Discussions. Proceedings of the Third IEEE International Conference on Social Computing, Boston, (forthcoming Oct. 2011).
• Kaiser, C.; Entscheidungsunterstützung zur Meinungsbeeinflussung in Webcommunitys In: HMD 280 (4) - Schwerpunkt Communitys im Web 2011.
• Kaiser, C.; Schlick, S.; Bodendorf, F.; Warning system for online market research - Identifying critical situations in online opinion formation In: Knowledge-based Systems 24(6) 2011.
• Kaiser, C.; Kröckel,  J.; Bodendorf, F.; Analyzing Opinion Formation in Online Social Networks - Mining Services for Online Market Research In: Proceedings of the 2011 Annual SRII Global Conference, IEEE, San Jose 2011.
• Kaiser, C.; Schlick, S; Frühwarnsystem zur Identifikation kritischer Situationen der Meinungsbildung im Web 2.0. In: Tagungsband Wirtschaftsinformatik 2011, Zürich 2011.
• Kaiser, C; Schlick, S.; Bodendorf, F.: Discovering Critical Situations in Online Social Networks - A Neuro Fuzzy Approach to Alert Marketing Managers. In: Proceedings of the 2nd International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, Valencia, 2010.
• Kaiser, C.; Bodendorf, F.: Analyse der Meinungsentwicklung in Online-Foren - Konzept und Fallstudie. In: K Meißner, M. Engelien (Hrsg.): Virtual Enterprises, Communities & Social Networks, Tagungsband Gemeinschaften in Neuen Medien 2010. Verlag der Wissenschaften, Dresden 2010, S. 271-280.
• Kaiser, C.; Kröckel, J.; Bodendorf, F.: Ant-based Simulation of Opinion Spreading in Online Social Networks. In: Proceddings of the 2010 IEEE / WIC / ACM International Joint Conference on Web Intelligence and Intelligent Agent Technologies. Toronto 2010.
• Kaiser, C.; Kröckel, J.: Meinungsanalyse in Onlinenetzwerken mittels Schwarmintelligenz. In: InformatikSpektrum, Springer, 2010.
• Bodendorf, F.; Kaiser, C.: Detecting Opinion Leaders and Trends in Online Communities. In: Proceedings of the Fourth International Conference on Digital Society. St. Maarten 2010.
• Kaiser, C.; Kröckel, J.; Bodendorf, F.: Swarm Intelligence for Analyzing Opinions in Online Communities. In: Proceedings of the Forty-Third Annual Hawaii Internationsal Conference on System Sciences. Kauai, Hawaii 2010.
• Bodendorf, F.; Kaiser, C.: Mining Customer Opinions on the Internet - A Case Study in the Automotive Industry. In: Proceedings of the International Conference on Knowledge Discovery and Data Mining. Puhket 2010.
• Kaiser, C.: Combinig Text Mining and Data Mining For Gaining Valuable Knowledge from Online Reviews. In: Pedro Isaías (Hrsg.): IADIS International Journal on WWW/Internet 6 (2009) 2, S. 63-78.
• Bodendorf, F.; Kaiser, C.: Detecting Opinion Leaders and Trends in Online Social Networks. In: Proceedings of the 2nd Workshop on Social Web Search and Mining. Hong Kong 2009.
• Kaiser, C.; Bodendorf, F.: Opinion and Relationship Mining in Online Forums. In: Proceedings of the 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology. Milan, Italy 2009, S. 128-131.
• Kaiser, C.: Opinion Mining im Web 2.0 - Konzept und Fallbeispiel. In: M. Knoll, A. Meier (Hrsg.): HMD Praxis der Wirtschaftsinformatik - Schwerpunkt Web & Data Mining (2009) 268.
• Kaiser, C.: Analyse von Meinungen in sozialen Netzwerken des Web 2.0. In: Hansen H. R., Karagiannis D., Fill H.-G. (Hrsg.): Business Services: Konzepte, Technologien, Anwendungen. 9. Internationale Tagung Wirtschaftsinformatik. Österreichische Computer Gesellschaft, Wien 2009.
• Kaiser, C.; Tirwana, B.; Bodendorf, F.: Briding the Gap between Qualitative and Quantitative Analysis of Opinion Forums. In: Proceedings of the 2008 IEEE/WIC/ACM International Conference on Web Intelligence. Sydney 2008.
• Kaiser, C.: Mining Customer Experience on the Web 2.0. In: P. Isaías; M. B. Nunes; D. Ifenthaler (Hrsg.): Proceedings of the IADIS International Conference WWW/Internet 2008. IADIS Press, Freiburg 2008, S. 3-11.
• Bodendorf, F.; Kaiser, C.: Case Based Reasoning. In: K. Kurbel; J. Becker; N. Gronau;E. Sinz; L. Suhl (Hrsg.): Enzyklopädie der Wirtschaftsinformatik - Online Lexikon. Oldenburg Wissenschaftsverlag, 2008.
• Kaiser, C.: Produkt-Mining im Web 2.0. In: Bichler, M; Hess, T.; Krcmar, H., Lechner, W.; Matthes, F.; Picot, A.; Speitkamp, B.; Petra W. (Hrsg.): Proceedings der Mulitkonferenz Wirtschaftsinformatik 2008. GITO-Verlag, Berlin 2008, S. 229-240.
• Bodendorf, F.; Kaiser, C.: Fallbasierte Individualisierung von Geschäftsprozessen und E-Services. In: WISU 37 (2008) 3.
• Schicker, G.; Kaiser, C.; Bodendorf, F.: Process and E-Service Customization - For Coordination in Healthcare Networks. In: Azevedo, L.; Londral, A. R. (Hrsg.): Proceedings of the First International Conference on Health Informatics. Funchal, Madeira 2008, S. 161-166.
• Kaiser, C.: Case Based Reasoning for Customizing Treatment Processes. In: Lazakidou, A. A.; Siassiakos, K. M. (Hrsg.): Handbook of Research on Distributed Medical Informatics and E-Health. Medical Information Science Reference, USA 2008.
• Schicker, G.; Kaiser, C.; Bodendorf, F.: Individualisierung von Prozessen und E-Services mithilfe von Case Based Reasoning. In: Oberweis, A.; Weinhardt, C.; Gimpel, H.; Koschmider, A.; Pankratius, V.; Schnizler, B. (Hrsg.): eOrganisation: Service-, Prozess-, Market-Engineering, 8. Internationale Tagung Wirtschaftsinformatik. Universitätsverlag Karlsruhe, Karlsruhe 2007, S. 713-730.
• Schertler, M.; Kaiser, C.: Der Einsatz von KI-Techniken zur Bewertung von Wissenskapital. Arbeitspapier Wirtschaftsinformatik II Nr. 03/2005, Universität Erlangen-Nürnberg. Nürnberg 2005.

Interviews

• Manta, C.: Analyse von Blogs und Social Networks - Interview mit Carolin Kaiser. In: Computerwoche, August 2009.
• Hoffmann, D.: Text Mining wird salonfähig - Wie Text Mining die Analyse moderner sozialer Netzwerke revolutionieren kann. In: IT Director, November 2009.

Diploma Theses and Result Papers

• F.Bernhard: Analyse von Zusammenhägen zwischen Meinungen im Web 2.0 und Unternehmenserfolgen. Diploma Thesis 2010.
• A. Schwingenschlögl: Opinion Mining im Web 2.0 zur Unterstützung des Sponsorship Managements. Diploma Thesis 2010.
• A. Gerhards: What-if-Analyse der Meinugnsbildung in Sozialen Netzwerken. Project 2010.
• A. Piazza: Entwicklung eines Systems zur Simulation der Meinungsentwicklung in Online Communities mittels Schwarmintelligenz. Master Thesis 2010.
• M. Bertram, M. Drechsler, K. Weigel, S. Volpe: Meinungsmanipulation im Web 2.0. Project 2010.
• K. Thomas, M. Czerwinka, V. Schaffranitz: Einstellungsbildung und -beeinflussung in Internetforen. Project 2010.
• A. Hamatschek, J. Rupp, T. Braun, V. Müller: Analyse der Schwarmintelligenz in Online-Foren. Project 2010.
• F. Zach: Prognose von Kundenmeinungen im Web 2.0 mittels Schwarmintelligenz. Project 2009.
• S. Schlick: Frühwarnsystem zur Identifikation entscheidungsrelevanter Situationen im Web 2.0. Diplomarbeit 2009.
• S. Winter: Einflussfaktoren der Meinungsbildung in sozialen Netzwerken des Web 2.0. Diploma Thesis 2009.
• C. Schuster, S. Cercan: Text Mining - Vergleich der Meinungsbewertung zwischen Mensch-Mensch und Mensch-Maschine im Bereich Consumer Electronics. Project 2008.
• A. Schwingenschlögel, F. Bernhard, C. Weinmann: Text Mining - Vergleich der Meinungsbewertung zwischen Mensch-Mensch und Mensch-Maschine in der Pharmabranche. Project 2009.
• C. Tandja, M. Aigner: Text Mining - Vergleich der Meinungsbewertung zwischen Mensch-Mensch und Mensch-Maschine in der Sportartikelbranche. Project 2009.
• C. Raschdorf, C. Schlagenhaufer: Text Mining - Vergleich der Meinungsbewertung zwischen Mensch-Mensch und Mensch-Maschine in der Automobilbranche. Project 2009.
• C. Schweiger: Link Mining als Methode zur Analyse von Kundennetzwerken. Thesis 2008.
• M. Sonntagbauer: Einsatzpotentiale des Web 2.0 für das Beziehungsmanagement. Diploma Thesis 2008.
• S. Frank: Schwarmintelligenz als Modell der Meinungsbildung im Web 2.0. Thesis 2008.
• J. Kröckel: Analyse der Meinungsentwicklung mit Hilfe von Methoden der Schwarmintelligenz. Diploma Thesis 2008.
• B. Berberich: Konzeption und prototypische Implementierung eines Systems zur dynamischen Analyse von Meinungen in Netzwerken des Web 2.0. Diploma Thesis 2008.
• T. Orgis: Konzeption und prototypische Implementierung eines Opinion Mining Systems. Thesis 2008.
• A. Schwingenschlögl, Florian Bernhard: Key Figures for Analysing Opinions in Social Networks. Project 2008.
• T. Birinda: Software for Analysing Opinions in Social Networks. Project 2008.
• R. Abraham: Konzeption und prototypische Implementierung eines Systems zur Extraktion von sozialen Beziehungen in Netzwerken des Web 2.0. Diploma Thesis 2008.
• K. Schröder: Konzeption eines Systems zur Analyse von Kundenmeinungen im Web 2.0. Diploma Thesis 2008.
• S. Jachtmann: Aufdeckung von Kundenbindungspotentialen im Projektlebenszyklus der Baumarktbranche mithilfe von Text Mining. Diploma Thesis 2008.
• A. Crummenauer: Text-Mining als Instrument der Online-Marktforschung. Diploma Thesis 2008.
• T. Göhner: Konzeption eines Systems zur Analyse geäußerter Arzneimittelwirkungen im Web 2.0. Diploma Thesis 2008.
• M. Weinmann: Diffusion von Meinungen in sozialen Netzwerken des Web 2.0. Diploma Thesis 2008.
• N. Spassova: Quantitative und qualitative Erfolgsgrößen von Web 2.0 Anwendungen auf Community Webseiten. Diploma Thesis 2008.
• J. Brückl: Einsatz von Text Mining zur Analyse von Produktbewertungen im Web 2.0 anhand eines prototypisch implementierten Systems. Diploma Thesis 2008.
• J. Kröckel: Entwicklung eines prototypischen Systems zur Analyse von Börsentrends mithilfe von Text Mining im Web 2.0. Studienarbeit 2008.
• A. Crummenauer, A. Egerer, J. Hetzenecker, S. Schlick: Intelligent Services - Data Mining am Airport. Project 2007.
• I. Hristeva, N. D. Ninh, M. Seegel: Meinungsanalyse von Produkten und ihren Eigenschaften im Web 2.0. Project 2007.
• M. Enders, S. Frank, C. Schweiger: Wissensmanagement und Semantic Web. Project 2007.
• B. Niemann, C. Oumard: Entwicklung Tutoriumkonzepts für den Einsatz von Social Software Systems in der universitäre Lehre. Project 2006.
• C. Kroll, J. Brückl, J. Forsbach, J. Bosmann, F. Mattes: Entwicklung eines Social Software Systems für die universitären Lehre. Project 2006.