|
Title:
|
A NEW HYBRID METHOD FOR DATABASE SELECTION IN MULTI-DATABASE MINING |
|
Author(s):
|
Aidin Davaran, Hassan Rashid |
|
ISBN:
|
978-972-8939-23-6 |
|
Editors:
|
António Palma dos Reis and Ajith P. Abraham |
|
Year:
|
2010 |
|
Edition:
|
Single |
|
Keywords:
|
Multi-database mining, database selection, distributed knowledge, relevant databases. |
|
Type:
|
Short Paper |
|
First Page:
|
153 |
|
Last Page:
|
156 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Multi-database mining refers to data analysis process in several dependent databases and finding a useful knowledge supported by most of these databases. In this kind of data mining, recognizing and selecting a database related to the desired data mining application is one the most important criteria to decrease cost of search. At first, this paper considers two of multi-database mining algorithms (classification algorithm and identifying of relevant databases algorithm). Then, a hybrid method including algorithm of the best classification and algorithm of the relevant databases is presented. The hybrid method offered in this paper, reduces time complexity of selecting relevant databases, obtains the best classification from among several databases and more exactly offers databases of a class which has the most relationship with application of the desired mining. Empirical outcomes resulted from simulation the hybrid method demonstrates that the database selection for multi-database mining in the proposed method is more effective and accurate compared with the two previous algorithms. |
|
|
|
|
|
|
