建筑设计项目任务书的评价指标提取方法研究

Research on Evaluation Index System of Design Brief

建筑设计任务书是建设项目必不可少的重要设计依据,其合理性直接影响着建筑创作的过程和最终设计方案的质量;然而,在当今的实践中,任务书所提出的设计依据与设计要求,往往过于随意草率,甚至缺乏理性。建筑策划是使用科学方式探讨任务书编制的一套理论与方法体系,在数据科学发展的今天,应该结合使用基于数据的方法,尝试针对任务书进行评价以实现建筑策划的自评,从而出具更加科学的、理性的任务书。本文以112份真实的任务书作为样本,基于任务书样本的数据类型,综合利用文本处理与数据挖掘技术,剖析任务书的一般内容构成与特征,同时识别出个体样本的特异性元素,以风险为导向归纳出任务书的评价指标,以期建立起一个初步的任务书评价体系。

Design brief is a critical basis document to architecture project.Its rationality highly affects design process and final scheme.However,conditions and requirements listed in design briefs are often simple and unprofessional in practice.Architectural programming is of theories and methods that help architects scientifically compile design briefs.As the field of data science is rapidly developing nowadays,it is promising to try data analysis methods on brief evaluation.It would be a new niche as well as self-examination of architectural programming,providing more scientific briefs.In this paper,we collect 112 real briefs as samples.Based on text and numerical data of these samples,by using techniques such as text mining,m achine learning and clustering,we study on the composition and feature of a general design brief;on the other hand,we also analysis those specifics and particularity of every single brief,locating potential risk items.As a conclusion,we try to give a design brief evaluation index system and establish a preliminary structure of it.