期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
Concept evolution in architectural design: an octonary framework 被引量:2
1
作者 Buthayna Eilouti 《Frontiers of Architectural Research》 CSCD 2018年第2期180-196,共17页
A framework that integrates the fragmented elements of concept derivation, processing, and translation is developed and discussed. This framework aims to consolidate the divergent components of design conception into ... A framework that integrates the fragmented elements of concept derivation, processing, and translation is developed and discussed. This framework aims to consolidate the divergent components of design conception into a comprehensive system that facilitates design conceptualization, initiation, and integration of various concept evolution phases, components, layers, and aspects. Accordingly, the framework describes a comprehensive approach based on a series of alternating divergence/convergence cycles. Subsequently, the approach encom- passes a concept evolution process, derivation methods, aggregation/segregation technique, translation channels, and development layers. Each of the major divergent phases of the framework consists of eight parts. Furthermore, the skills and tasks associated with framework implementation are mapped into the main processing phases. The scope of this study is architectural design in higher education and practice. To demonstrate its applicability, the framework is implemented and illustrated by a case study. Reflections about its implementation and limitations are reported and discussed. 展开更多
关键词 Design ideation concept generation design systematiza-tion idea concretization concept objectifica-tion concept consolidation design management
原文传递
中医脉诊信号的无监督聚类分析研究 被引量:10
2
作者 冯冰 李绍滋 《智能系统学报》 CSCD 北大核心 2018年第4期564-570,共7页
随着中医客观化工作的推进,脉诊技术也越来越走向客观化和仪器化。然而,如何对仪器所检测和收集到的信息进行解读,却还是回到了原来脉诊诊断主观化的问题上。因为传统的机器学习方法,依赖于对大量的脉诊数据进行标注。但是在临床诊断和... 随着中医客观化工作的推进,脉诊技术也越来越走向客观化和仪器化。然而,如何对仪器所检测和收集到的信息进行解读,却还是回到了原来脉诊诊断主观化的问题上。因为传统的机器学习方法,依赖于对大量的脉诊数据进行标注。但是在临床诊断和教学中,医生与医生之间对于脉象的体会不同,会导致他们对病人脉象的区分标注不同。在对比了多种特征提取方法和聚类方案之后,提出了一个较好的无监督脉诊客观化方法,在双树复小波变换(DTCWT)对数据进行预处理的基础上,以梅尔倒谱系数(MFCC)进行特征提取,在中医专家对数据进行标注之前,先根据信号的特征,使用Fuzzy c-means(FCM)聚类算法进行粗线条的分类,使得在此基础之上,可以开展进一步的细化分类研究。实验结果表明:该方法可取得较好的分类效果,为中医脉诊提供了进一步客观化的依据。 展开更多
关键词 脉诊 机器学习 无监督学习 聚类分析 双树复小波变换 中医客观化 梅尔倒谱系数 模糊C均值聚类
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部