CAJ | 학술논문

演化趋势检测能提前获取话题的演变轨迹，有效提高决策者对于话题舆情处理的主动性和预判性。针对该需求，该文深入研究新闻话题的演变过程，利用K-means算法对专题新闻报道集进行聚类，得到新闻话题演化阶段的不同类簇，并且采用Text Rank方法提取各个新闻话题演化阶段的有效关键词，然后基于时间信息，整理生成话题演化趋势。该文以百度百科相关词条的客观描述作为评判标准进行算法评测，实验结果表明本文算法具有较高的准确性。
연화추세검측능제전획취화제적연변궤적，유효제고결책자대우화제여정처리적주동성화예판성。침대해수구，해문심입연구신문화제적연변과정，이용K-means산법대전제신문보도집진행취류，득도신문화제연화계단적불동류족，병차채용Text Rank방법제취각개신문화제연화계단적유효관건사，연후기우시간신식，정리생성화제연화추세。해문이백도백과상관사조적객관묘술작위평판표준진행산법평측，실험결과표명본문산법구유교고적준학성。
The evolutionary trend detection can obtain the topic evolution track in advance, and it can effectively improve the initia?tive and anticipation of the decision for the resolution of the public opinion on the news topic. According to the requirement, this pa?per studies the evolution of news topic firstly, using k-means algorithm for clustering of special news and getting different clusters for different stages of news topic evolution. And then we extract the keywords existing in various stages of news topic evolution with Text Rank model to represent the topic evolution trend based on temporal information in news pages. Finally, this paper takes the objective description of the relevant entries of Baidu encyclopedia as the evaluation criteria for the assessment of the algorithm, and the experimental results show that the algorithm is of high precision.