具一致性與可調性的摘錄型生物資料探勘仲介者

Abstract

在進入後基因時代的今天，網際網路上的生物資料數量正以驚人的速度在成長著。多種異質性的資料來源､不同的探勘應用程式與知識分析師的存在，導致語句與語意上的異質性現象層出不窮。本論文提出一個具一致性與可調性的摘要型生物資料探勘仲介者，用以整合與克服形式與意義上的異質性現象。該生物資料探勘仲介者提供一個聯邦機制的架構，藉以增強系統中以回饋模式為基礎之意境導向快取機制的功能。而透過模擬實驗也顯示了，那些儲存在生物資料探勘仲介者中快取資料庫的摘要型生物資料，經由生物資料探勘仲介者所提供之以回饋模式為基礎的意境導向快取機制，得以維持並確保其意義性及有效性。

Now we have stepped into the post-genetic age. The volume of biological data in the Internet is increasing in an extremely high speed. With multiple heterogeneous data sources, different mining applications and knowledge analysts, it results in syntactic and semantic heterogeneities. The unified, adjustable and extractable biological mining-broker is proposed to integrate and overcome heterogeneities of form and meaning in this thesis. It provides the federated model to enhance the feedback-based meaningful cache mechanism. The experimental simulation also shows that the feedback-based meaningful cache mechanism ensures and maintains the meaningfulness and usefulness of extraction cache information in the biological mining-broker.