耦合文本挖掘与高维排序权重法的综合指数算法及中国省级发展评价应用案例  

Composite index algorithm coupling text mining and high dimensional ranked weights method and application case in provincial development evaluation in China

作　　者：王元慧 宋长青 王翔宇 高培超 WANG Yuanhui;SONG Changqing;WANG Xiangyu;GAO Peichao(State Key Laboratory of Earth Surface Processes and Resource Ecology,Beijing Normal University,Beijing 100875,China;Center for Geodata and Analysis,Faculty of Geographical Science,Beijing Normal University,Beijing 100875,China)

机构地区：[1]北京师范大学地表过程与资源生态国家重点实验室,北京100875 [2]北京师范大学地理科学学部地理数据与应用分析中心,北京100875

出　　处：《地理研究》2023年第10期2556-2573,共18页Geographical Research

基　　金：国家自然科学基金重点项目(42230106);国家自然科学基金项目(42271418)。

摘　　要：科学实用的综合指数算法对监测区域发展状况有重要意义。综合指数算法的关键在于权重量化,其中以主观权重的量化最为困难。少数个体的主观权重量化难以进行精确的数值化且缺乏代表性;群体的主观权重量化中存在意见收集过程繁杂的困难。本文提出耦合文本挖掘与高维排序权重法的综合指数算法,其中文本挖掘技术能够基于文本对各维度进行权重数值排序提取、解决主观意见获取与量化难题,高维排序权重法能利用权重数值排序计算综合指数、降低权重量化要求。为证明算法的可用性,基于政策文本挖掘政府决策者主观意见,逐年量化了符合政策导向和时代特色的维度权重排序并分析了中国2000—2019年省级区域发展质量水平。结果表明,新算法评价结果合理、可用性强。中国省级区域发展质量整体水平前、中期平稳上升,后期波动上升;东部沿海省份形成高水平聚集区;随着中部和西南省份的崛起,东西发展不均衡问题有所缓解。本文耦合算法的应用不局限于案例中的区域发展,可在大数据时代的数据红利支持下进行高效、低成本的指数评价和决策支持。Scientific and useful composite index algorithms are of great significance for monitoring regional development situation.The key of composite index algorithms is the quantification of weights,among which the quantification of subjective weights is most difficult.In the quantification of subjective weights based on only few experts,it is difficult to accurately quantify the subjectivity,and weights are not representative enough.In the quantification of subjective weights based on a group of experts,there exist difficulties in the complex process of collecting subjective opinions.This study proposes a composite index algorithm that couples the text mining technology and the high-dimensional ranked weights method.The high-dimensional ranked weights method could utilize ranks of weight values of basic indexes or dimensions to calculate comprehensive indexes and reduce the requirement of weight quantification.Text mining technology could automatically extract ranks of weight values based on massive texts and solve the problem of obtaining and quantifying subjective opinions of the group.To prove the usefulness of this algorithm,we mined subjective opinions of government decision-makers based on policy texts so as to annually quantify ranks of weights for dimensions reflecting policy orientations and era characteristics and analyzed development qualities of China's provinces in 2000-2019.Results show that the proposed algorithm is strongly useful,and the evaluation results are reasonable.The overall level of China's provincial development quality increased steadily in the early and middle periods and increased with fluctuations in the later period;eastern coastal provinces form high-level clusters;with the rise of some central and southwestern provinces,the imbalance between the eastern and the western regions has been alleviated to some extent.The application scope of the coupling algorithm in this study is not limited to regional development evaluated in the example.Making full use of data benefits of this period of

关 键 词：综合指数 文本挖掘 高维排序权重法 高质量发展 政府决策 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术] TP311.13[自动化与计算机技术—计算机科学与技术] F127[经济管理—世界经济]