資料發掘研究重點在於幫助使用者於眾多現存資料中發掘出隱含於其內而先前未知的可能有用資料。目前有三大主要研究派別：(1)類神經網路(2)歸納學習方法論(3)統計方法。由於本研究之研究目的在於加入先前知識於資料發掘過程中，因此選用歸納學習方法論。歸納學習方法其內又可分為樹狀分類法，關聯分析法及概念樹導向歸納學習法，由於採概念樹導向歸納學習法所能處理的資料發掘問題種類較完整，其它二種歸納學習方法均著重於某一特定種類的資料發掘問題處理，因此，本研究針對概念樹導向歸納學習法做研究基礎，探討先前知識的種類及其運用方式，以期能增加資料發掘後的意義性。
首先從文獻中了解目前資料發掘領域的研究現況，從而由擴充先前知識的角度切入，利用企業法則、實體層次之一般化、集合化、聚集化等抽象化觀念、延伸之資料字典及經驗法則等先前知識得出更合適的資料以供資料發掘，並對於概念樹導向歸納學習法做適當的修改，提出研究架構。再以假想的學校資料庫，發展出一套雛形系統，驗証本架構的可行性。最後提出進一步的研究建議，以供後續研究參考。 / The research objective of data mining is to help users find previous unknown and maybe usable information from database. There are three ways to do this:(l)neutral network (2)inductive learning (3)statistics. Inductive Learning has three different ways: learning by decision trees, association rules and using concept trees.
Because concept trees approach to inductive learning can solve more kinds of problem, the other two ways just can solve one kind of problem, we choose using concept trees to be our foundation of this research. At the same time, we explore and discuss serveral kinds of prior knowledge and their applications. We hope that it can increase the semantics of mining results.
This thesis, first surveys previous research in data mining and discuss the prior knowledge that they included. Then, we propose our idea of extending and using prior knowledge including data abstractions (generalization, association and aggregation) in the extended entity-relationship model, bussiness rule, extended data dictionary and heuristics, in order to assist the process of data mining. A prototype is reported to prove our research architecture. Finally, some sugestion are given to future research.
| Identifer | oai:union.ndltd.org:CHENGCHI/A2010000736 |
| Creators | 林幸怡 |
| Publisher | 國立政治大學 |
| Source Sets | National Chengchi University Libraries |
| Language | 中文 |
| Detected Language | English |
| Type | text |
| Rights | Copyright © nccu library on behalf of the copyright holders |
Page generated in 0.0018 seconds