A Neuro Fuzzy System for Knowledge Discovery of Incomplete Construction Data
Autor: | Han-Wen Lin, 林涵文 |
---|---|
Rok vydání: | 2004 |
Druh dokumentu: | 學位論文 ; thesis |
Popis: | 92 Construction management (CM) is a discipline that replies heavily on the skills and experiences of personnel, and is very suitable for application of data mining (DM) techniques to acquire the domain knowledge for future use. Unfortunately, the DM application in CM is facing a severe problem of data incompleteness. There has been no exiting method for mining of incomplete data thus far. This research proposes a Variable-attribute Fuzzy Adaptive Learning Control Network (VaFALCON) approach to tackle data incompleteness in historical construct databases. At first, the problem of data incompleteness is defined. Secondly, computational algorithm of VaFALCON is developed step by step. Then, experiments are designed to verified the proposed VaFALCON method. It is found from the demonstrated examples that the proposed VaFALCON can improve the system accuracy from 9% up to 26% while the overall incompleteness ranges from 5% to 20%. Finally, findings of research are concluded and future works are recommended. |
Databáze: | Networked Digital Library of Theses & Dissertations |
Externí odkaz: |