完整後設資料紀錄
DC 欄位語言
dc.contributor.author王振生en_US
dc.contributor.authorWang, Chen-Shengen_US
dc.contributor.author蔡銘箴en_US
dc.contributor.authorTsai, Min-Jenen_US
dc.date.accessioned2014-12-12T01:22:26Z-
dc.date.available2014-12-12T01:22:26Z-
dc.date.issued2009en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#GT079234810en_US
dc.identifier.urihttp://hdl.handle.net/11536/40442-
dc.description.abstract網際網路蓬勃發展,把資訊系統的使用範圍拓展到整個供應鏈。環顧整個資訊系統發展歷程,其在決策支援上所扮演的角色也越來越重,系統使用者對資料背後所代表的資訊甚至智慧的需求也越來越高,例如在營運資源分配最佳化的前提下,如何從供應鏈涵蓋的採購、生產、配送及銷售等四個資訊系統的資料中,融合出產銷規劃的決策支援資訊,便是一個典型的例子,因此如何幫助使用者找到決策支援資訊,便成為資訊系統發展的關鍵。過往,這些決策支援資訊大都由單一公司的單獨系統所提供,但網際網路普及後,已使不同公司不同系統間所產生的決策支援資訊大量增加,但這些資訊可能因建立目的不同而互相抵觸,因此如何尋求這些資訊的融合便有其必要性。本研究修改資料融合處理架構,以應用到現今資料多樣化的環境,提出一個資訊融合架構,此架構使用柔性計算的模糊理論及社會科學在共識理論的研究,將其應用在網路服務為基礎的分散式運算及影像來源相機辨識之資訊融合部份,當應用在網路服務為基礎的分散式運算時,所提的架構融合不同的運算資源資訊,產生一組工作分派的指令,使得整體運算執行時間因而減少5.2%,穩定性增加76.7%;如應用於影像來源相機辨識時,所提的架構融合5個不同特徵選取演算法的結果,產生一組最佳特徵子集合,將辨識所需的特徵數目從34個減少為20個,但整體辨識率從88.85% 改善至90.68%,運算時間減少43%, 從這些實驗結果驗證了所提架構的有效性。zh_TW
dc.description.abstractThe boom of the Internet has extended the scope of information systems to cover the whole enterprise supply chain. Through the evolving process of information systems, the more important role a decision support system plays, the more demands to mine the intelligence behind information in the enterprise database. For example, to get optimized resources allocation, it is a typical application to obtain decision support information on production-distribution planning by fusing data from purchasing, production, distribution and sales information systems in a supply chain. Therefore, how to help users find decision support information has been the key point during the development of information systems. In the past, decision support information was created in a single system but the Internet is raising the need to generate such information from different systems in or among the enterprises. However, the decision support information from the different systems may conflict because the perspective in each system is unlike. Hence, it is necessary to find the consensus among the decision support information from different systems. To explore the way to get consensus, the information fusion model is proposed in this study by making modification on multisensory fusion model in order to apply the model in the environment having various formats of information. This model is applied in the information confusing process of web-service based distributed computing and image’s source camera identification by using fuzzy theory of soft computing and consensus theory which is well researched in the sociology. When it is applied in web-service based distributed computing by fusing resource information from various computing units, it generates a queue of tasks assignment which results in decreasing 5.2% execution time and increasing 76.7% stability. If the model is utilized in image’s source camera identification, it generates an optimum feature subset by fusing the result from 5 feature selection algorithms. Although this optimum subset contains only 20 features selected from 34 ones in the full set, it lifts not only the accuracy rate from 88.85% to 90.68% but also reduces 43% execution time. These experimental results justify the proposed model proposed in this study.en_US
dc.language.isozh_TWen_US
dc.subject資訊融合zh_TW
dc.subject支持向量機zh_TW
dc.subject模糊偏好關係zh_TW
dc.subject意見融合zh_TW
dc.subjectInformation Confusionen_US
dc.subjectSupport Vector Machineen_US
dc.subjectFuzzy Preference Relationen_US
dc.subjectOpinion Fusionen_US
dc.title模糊資訊融合機制在分散式運算及影像來源相機辨識之應用zh_TW
dc.titleApplying Fuzzy Information Confusion for Distributed Computing and Image's Source Camera Identificationen_US
dc.typeThesisen_US
dc.contributor.department資訊管理研究所zh_TW
顯示於類別:畢業論文