国际新闻报道主题模型分析及解释性理论探思

The TM Analysis of International News Reports and the Explanatory Theories

意义与意图共生共存于文本之中,解读意义和透视意图一直是国际新闻报道研究的重要维度。基于海量文本数据的计算机主题模型分析,能够在字符关联中显现意义,在语义结构中洞见意图,正在成为国际传播研究的新方法。然而,当前引入隐含主题挖掘技术的人文社科研究存在一个共同的瓶颈,即多数囿于方法创新,缺乏学科理论阐释。以国际新闻报道主题模型研究为例,在意义解读层面,具有结构分析能力的"框架理论"和拥有内容深究优势的"语境理论"在隐含主题分析上呈互补之态;而在意图透视层面,分析显在意图的"国家利益理论"和揭示隐藏意图的"镜像理论"在隐含主题分析时可互为表里。单纯的隐含主题挖掘及其文本关联聚类的主题模型数据结果,只有结合能够与之互洽的学科解释性理论才能实现其学术和应用价值,人文社科其它研究领域亦同此理。

Meaning and intention coexist in the text, and the interpretation of meaning and the perception of intentions have always been important dimensions in the study of international news reports. The analysis of topic modeling based on massive text data can show meaning in character association and reveals intentions in a semantic structure, which is becoming a new method of international communication research. However, a common bottleneck in the studies of humanities and social sciences that introduce LDA mining technology is the mismatching of the creative technical methods and the insufficient professional depth. Taking the study of the subject model of international news reports as an example, at the level of meaning interpretation, the"framing theory"with an ability in structural analysis and the"context theory"with an advantage in content depth are complementary in the analysis of latent topics, while at the level of intention perception, the"national interests theory"analyzing apparent intentions and the"mirror theory"revealing hidden intentions play exterior and interior roles in the analysis of latent topics respectively. The results of TM analysis in latent topics mining and text association clustering canachieve their academic and practical value only in combination with the explanatory theory, and the same is true of other research fields of humanities and social sciences.