Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
This paper presents the results of experiments in which we tested different kinds of features for retrieval of Chinese opinionated texts. We assume that the task of retrieval of opinionated texts (OIR) can be regarded as a subtask of general IR, but with some distinct features. The experiments showed that the best results were obtained from the combination of character-based processing, dictionary look up (maximum matching) and a negation check. | Kinds of Features for Chinese Opinionated Information Retrieval Taras Zagibalov Department of Informatics University of Sussex United Kingdom T.Zagibalov@sussex.ac.uk Abstract This paper presents the results of experiments in which we tested different kinds of features for retrieval of Chinese opinionated texts. We assume that the task of retrieval of opinionated texts OIR can be regarded as a subtask of general IR but with some distinct features. The experiments showed that the best results were obtained from the combination of character-based processing dictionary look up maximum matching and a negation check. 1 Introduction The extraction of opinionated information has recently become an important research topic. Business and governmental institutions often need to have information about how their products or actions are perceived by people. Individuals may be interested in other people s opinions on various topics ranging from political events to consumer products. At the same time globalization has made the whole world smaller and a notion of the world as a global village does not surprise people nowadays. In this context we assume information in Chinese to be of particular interest as the Chinese world the mainland China Taiwan Hong Kong Singapore and numerous Chinese communities all over the world is getting more and more influential over the world economy and politics. We therefore believe that a system capable of providing access to opinionated information in other languages especially in Chinese might be of great use for individuals as well as for institutions in volved in international trade or international relations. The sentiment classification experiments presented in this paper were done in the context of Opinionated Information Retrieval which is planned to be a module in a Cross-Language Opinion Extraction system CLOE . The main goal of this system is to provide access to opinionated information on any topic ad-hoc in a language different to the .
