社会经济研究中加权建模的初步研究  

APPLICATION OF TRANSCENDENTAL KNOWLEDGE TO SOCIAL ECONOMY MODELING: THE CASE OF WEIGHTED MODELING

作　　者：吴殿廷[1] 周伟[1] 姜晔[1] 

机构地区：[1]北京师范大学地理学与遥感科学学院,北京100875

出　　处：《北京师范大学学报（自然科学版）》2008年第5期524-528,共5页Journal of Beijing Normal University(Natural Science)

基　　金：国家自然科学基金资助项目(90102013)

摘　　要：提高建模精度,充分利用已有的信息,努力挖掘先验知识的价值,对社会科学研究非常必要,加权建模是其重要途径.给出了社会科学研究中常见模型的加权建模计算方法,讨论了典型情况下权重系数的确定原则,结合我国各地区现代化与工业化、城镇化、知识化及经济发展的情况,验证了相关分析、逐步回归分析和回归选优模型加权建模的可行性和有效性.无论是自然科学还是社会科学,都应提倡加权建模,但在权重系数选择上则要慎重,努力避免人为偏见.Increasing modeling precision, fully exploring acquired information and transcendental knowledge, is necessary for social science research. Weighted modeling is an important approach in the respect. This paper presents methods of common models for weighted modeling in social science research. Confirming principle of weighted coefficient in representative eases including each regional modernization, industrialization, urbanization, intellectualization and economic development was discussed. Feasibility and validity of correlating analysis, stepwise regression analysis and regression selected model weighted modeling was confirmed. Weighted modeling should be used increasingly both in natural science and social science. But particular care must be taken when when choosing weighted coefficients, to avoid artificial biases.

关 键 词：先验知识 加权建模 社会科学 

分 类 号：O178[理学—数学] V212.1[理学—基础数学]