New data mining techniques to deal with both Static and dynamic databases

Dissertant

Nasir al-Din, Hibah H. O.

Thesis advisor

al-Ajluni, Naim M. M.

University

Arab Academy for Financial and Banking Sciences

Faculty

The Faculty of Information Systems and Technology

University Country

Jordan

Degree

Ph.D.

Degree Date

2007

English Abstract

-Dynamic data mining has many applications and can be considered a challenging research problem.

Typical applications, such as web searching, telephone services, network monitoring, and credit card purchases, all such applications are characterized by the need to mine continuously massive dynamic data to discover up-to-date patterns, which are invaluable for timely strategic decisions.

These new requirements call for the design of new mining methods to replace the traditional Algorithms, since conventional algorithms require data to be first stored and then processed off-line using complex algorithms that make several passes over the data.

Therefore, this dissertation presents the design of a fast and light mining algorithm for dynamic data.

The problem of noise represents a general problem in data mining; it creates new challenges when considering dynamic data insofar as adaptability becomes more difficult when the dynamic data contains noise.

To deal with noise and improve mining quality, the developed algorithm uses a general algorithm is used based on data dependency, it also exploits local data dependency between samples.

Moreover, the algorithm is highly adaptive through novel change detection process.

The algorithm developed in this dissertation deals with data collection ; it also detects changes that take place on selected and / or extracted data within its original data source, this change is reflected on the data extracted and stored in the temporary local database.

The algorithm also presents a new and efficient method developed to detect and selects records that have been changed and / or modified in the original data source, during and after a data mining run, regardless of the size of data available within the source database.

This algorithm uses a mathematical equation developed special for the purpose of generating summation value for each record.

The new updates and their summation values are stored in a dummy table in our local database.

Based on the summation values the exact record in the local database that needs updating (inset, update, delete) can be identified.

The proposed algorithm will deal the second algorithm used for the maintenance of the association rule produced by the data mining algorithm.

It uses the data stored in the dummy table to modify the association rule based on the actual update (insert, update, and delete) transaction The efficiency of the proposed algorithm is demonstrated through extensive experiments, both on synthetic and on real-life data.

Main Subjects

Information Technology and Computer Science

Topics

• Electronic searching
• Data mining

No. of Pages

164

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-304921