Sentiment Analysis: Extracting Decision-Relevant Knowledge from UGC

A novel approach for extracting decision-relevant knowledge from UGC is presented and different underlying data mining techniques concerning their accuracy in topic and sentiment detection of textual user reviews are compared.

Abstract

Electronically available user generated content (UGC) dramatically increased in recent years and constitutes a highly relevant information source not only for other customers but also for tourism suppliers. Customer needs and their perception of consumed products can be extracted from UGC and represent a valuable input to product enhancement and customer relationship management. A prerequisite to that end is an automatic extraction of decision-relevant knowledge from UGC with a sufficient quality. This paper presents a novel approach for extracting decision-relevant knowledge from UGC and compares different underlying data mining techniques concerning their accuracy in topic and sentiment detection of textual user reviews. The complete extraction process is implemented and evaluated in the context of the Swedish mountain tourism destination Åre.