Mining consumers’ opinions on the web

Andreas Auinger, Martin Fischer

Research output: Chapter in Book/Report/Conference proceedingsConference contribution

1 Citation (Scopus)


Comparing consumer’s opinion concerning own products and those of the competitors to find their strengths and weaknesses is a crucial activity for marketing specialists in the production industry to overcome the requirements of marketing intelligence and product benchmarking. Hence, web-forums, blogs and product review websites provide valuable findings and discussions that record the public voice. Therefore, a huge variety of opinions and commentary about consumer products is woven into the web, which offers a new opportunity for companies to understand and respond to the consumer by analyzing this raw feedback. This paper presents an approach that combines results and understand-ings from several procedures to encounter the challenge of opinion mining. The proposed architecture includes a wide variety of state-of-the-art text mining and natural language processing techniques. Fur-thermore, the key elements of applications for mining large volumes of textual data for marketing intelli-gence are reasoned: a suite of powerful mining and visualization technologies and an interactive analy-sis environment that allows for rapid generation and testing of hypothesis. The concluding results show that recent technologies look promising, but are still far away from a semantically correct text under-standing. Furthermore, this paper presents the results of a proof-of-concept of Text Mining with SPSS software. It is argued that SPSS Text Mining cannot meet the requirements to perform opinion mining as request by market research in the moment.
Original languageEnglish
Title of host publicationFH Science Day 2008, Linz
Publication statusPublished - 2008
EventScience Day 2008 - Linz, Austria
Duration: 6 Nov 20086 Nov 2008


WorkshopScience Day 2008


Dive into the research topics of 'Mining consumers’ opinions on the web'. Together they form a unique fingerprint.

Cite this