Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
Two psycholinguistic and psychophysical experiments show that in order to efficiently extract polarity of written texts such as customerreviews on the Internet, one should concentrate computational efforts on messages in the final position of the text. | Last but Definitely not Least On the Role of the Last Sentence in Automatic Polarity-Classification Israela Becker and Vered Aharonson AFEKA - Tel-Aviv Academic College of Engineering 218 Bney-Efraim Rd. Tel-Aviv 69107 Israel lsraelaB Vered @afeka.ac.il Abstract Two psycholinguistic and psychophysical experiments show that in order to efficiently extract polarity of written texts such as customerreviews on the Internet one should concentrate computational efforts on messages in the final position of the text. 1 Introduction The ever-growing field of polarity-classification of written texts may benefit greatly from linguistic insights and tools that will allow to efficiently and thus economically extract the polarity of written texts in particular online customer reviews. Many researchers interpret efficiently as using better computational methods to resolve the polarity of written texts. We suggest that text units should be handled with tools of discourse linguistics too in order to reveal where within texts their polarity is best manifested. Specifically we propose to focus on the last sentence of the given text in order to efficiently extract the polarity of the whole text. This will reduce computational costs as well as improve the quality of polarity detection and classification when large databases of text units are involved. This paper aims to provide psycholinguistic support to the hypothesis which psycholinguistic literature lacks that the last sentence of a customer review is a better predictor for the polarity of the whole review than other sentences in the review in order to be later used for automatic polarity-classification. Therefore we first briefly review the well-established structure of text units while comparing notions of topicextraction vs. our notion of polarityclassification. We then report the psycholinguistic experiments that we ran in order to support our prediction as to the role of the last sentence in polarity manifestation. Finally we
