TU Berlin

Database Systems and Information Management GroupData Programming Languages: History and Future

Logo FG DIMA-new  65px

Page Content

to Navigation

Short info

Candidate: Guilhelm Simons

Advisor: Johannes Kirschnick,  Peter Gerngroß (mayato GmbH)

Desired degree: Diploma

Abstract

This presentation gives voice to a methodology for measuring the impact of TV
media advertising using large-scale electronic Word of Mouth (eWOM) from
the German-language web. Many research studies have focused on English-
language eWOM and sentiment analysis. Little research quantitatively ana-
lyzes the impact of TV advertising in eWOM.

This paper addresses the problem of quantitative TV advertising analytics in
eWOM, investigating the potential causality linking TV advertising with brand
awareness and image in German-language eWOM. Our approach includes var-
ious information retrieval and natural language processing tasks: topic match-
ing; temporal information extraction and interpretation; sentiment analysis
linked with TV media analytics.

Based on a case study from a German pure-play start-up company, we ex-
tracted relevant web documents from the MIA project corpus, estimated their
respective publication date, analyzed the sentiment-loaded content, and linked
our results to media analytics to gain insights into the impact of TV adver-
tising in eWOM. Finally we address the limitations of our approach in both
the methodology and the analysis, and we give a brief outlook on prospective
future research areas in this matter.

Navigation

Quick Access

Schnellnavigation zur Seite über Nummerneingabe