面向可拓建筑设计的数据准备流程与方法

 为面向可拓建筑设计进行可拓数据挖掘, 以可拓建筑设计数据为对象, 探讨将其转化为高质量的结构数据的流程与方法。根据跨行业数据挖掘标准流程(CRISP-DM), 建立包括基元化表达、数据表设计、数据筛选、数据形式变换、变量标准化处理和变量维数约简6 个步骤的数据准备流程, 并根据建筑学专业特点, 结合可拓学、几何学、统计学理论, 设计各步骤的操作方法, 构建出完整的数据准备流程与方法。案例检验结果表明, 按照该流程及其操作方法, 可有效地将可拓建筑设计数据转化为统一格式、高信度、量化且可运算的结构数据。

Process and Methods of Data Preparation for Extension Architecture Design

The process and method of transforming extension architecture design data into structured data are investigated in order to execute extension data mining on the extension architecture design. According to the cross-industry standard process for data mining (CRISP-DM), a 6-stage data preparation process is built, which includes expression in basic-element form, data sheet design, data filtering, data form transform, standardization of variables, and variable dimension reduction. According to the characteristic of architecture, the methods for each stage are established by referring to extenics, geometry, statistics, etc. Finally, an integral data preparation system of the process and methods is constructed.