基于微博数据的城市形象评价方法

A City Image Evaluation Method Based on Weibo Data

现有城市形象评价方法大都依赖于评价者主观提出的指标体系,针对此现象,提出一套以微博数据为基础的城市形象评价方法,在一定程度上提升评价过程的客观性与可重复性。首先对获取到的城市数据预处理后,运用K-Means聚类算法得到城市形象的评价类别;其次采用LDA主题模型对有效类别总结主题、提取二级指标操作,利用情感分析技术计算目标城市情感得分即为城市形象评分;最后以武汉市和苏州市为例,利用该方法评价了城市形象。

Most of the existing urban image evaluation methods rely on the index system proposed by the evaluators subjectively.Aiming at this phenomenon, a set of urban image evaluation methods based on Weibo data is proposed to improve the objectivity and repeatability of the evaluation process to a certain extent.. First, after preprocessing the acquired city data, use the K-Means clustering algorithm to get the evaluation category of the city image;secondly, use the LDA topic model to summarize the topics of the effective categories, extract the secondary index operations, and use the sentiment analysis technology to calculate the target city sentiment The score is the city image score;finally, using this method to evaluate the city image, taking Wuhan and Suzhou as examples.