STARLET: Multi-document Summarization of Service and Product Reviews with Balanced Rating Distributions

Author

Fabbrizio, Giuseppe Di ; Aker, Ahmet ; Gaizauskas, Robert

Author_Institution

Dept. of Comput. Sci., Univ. of Sheffield, Sheffield, UK

fYear

2011

fDate

11-11 Dec. 2011

Firstpage

67

Lastpage

74

Abstract

Reviews about products and services are abundantly available online. However, selecting information relevant to a potential buyer involves a significant amount of time reading user´s reviews and weeding out comments unrelated to the important aspects of the reviewed entity. In this work, we present STARLET, a novel approach to multi-document summarization for evaluative text that considers the rating distribution as summarization feature to consistently preserve the overall opinion distribution expressed in the original reviews. We demonstrate how this method improves traditional summarization techniques and leads to more readable summaries.

Keywords

information retrieval; reviews; text analysis; STARLET; balanced rating distribution; evaluative text; multidocument summarization; opinion distribution; product review; service review; summarization feature; user reviews; Adaptation models; Data mining; Feature extraction; Measurement; Predictive models; Redundancy; Training; A* search; Summarization; evaluative text; multi-ratings prediction;

fLanguage

English

Publisher

ieee

Conference_Titel

Data Mining Workshops (ICDMW), 2011 IEEE 11th International Conference on

Conference_Location

Vancouver, BC

Print_ISBN

978-1-4673-0005-6

Type

conf

DOI

10.1109/ICDMW.2011.158

Filename

6137362