【目的/意義】從輿論對象的情感變化和關系變化展現輿情走向,為輿情監測和分析提供新的研究方法和研究角度。【方法/過程】文章提出輿論對象分析模型,利用依存句法分析,識別和抽取輿論對象-情感詞對,進行情感分析,并對輿論對象的關系網絡進行研究。【結果/結論】實驗結果表明,模型能有效識別主要輿論對象及其情感詞,直觀地展現網民對輿論對象隨時間演化的情感表達和關系認知。通過輿論對象情感變化與輿情事件發展的擬合,可為輿情監測、分析等相關研究提供新的研究視角。

【Purpose/significance】In order to describe and analyze the public event from the angle of public opinion objects.【Method/process】This paper puts forward a sentiment and relation network analysis model of public opinion objectsbased on the dependency parsing.【Result/conclusion】The results show that the model can effectively identify the objectsof the public opinion event and their sentiment words, and display the trend of public opinion from the change of the senti-ments and the change of the relations between those objects, and also, the results reveal some rules about how event goes.