Integration of useful links in distributed databases using decision tree classification

Abstract

Nowadays, distributed relational databases constitute a large part of information storage handled by a variety of users. The knowledge extraction from these databases has been studied massively during this last decade. However, the problem still present in the distributed data mining process is the communication cost between the different parts of the database located naturally in remote sites. We present in this paper a decision tree classification approach with a low cost communication strategy using a set of the most useful inter-base links for the classification task. Different experiments conducted on real datasets showed a significant reduction in communication costs and an accuracy almost identical to some traditional approaches.

SCATTER: Fully Automated Classification System across Multiple Databases

“…Moreover, our system performs better in finding the most useful joins across the data sources, thanks to the regression model used in predicting the link usefulness [4,5,6]. To perform the classification task, we use the decision tree classification algorithm that exploits the joins discovered automatically across the databases [4,5,6]. Experiments performed on five real databases were very satisfactory and show that the proposed system succeeded in achieving a fully automatic classification across multiple heterogeneous databases.…”

“…Putting all the data from the relevant databases into a single data set can destroy some important information that reflects the individuality of the different databases. 6. And the important limitation is the heterogeneity problem, where the aggregation of all the heterogeneous databases to obtain a whole single database could be simply an unfeasible solution.…”

“…To solve the heterogeneity problem, we partially follow the work of [3], where the author presented a framework that automatically identifies approximate foreign-key joins in the multiple heterogeneous databases. Moreover, our system performs better in finding the most useful joins across the data sources, thanks to the regression model used in predicting the link usefulness [4,5,6]. To perform the classification task, we use the decision tree classification algorithm that exploits the joins discovered automatically across the databases [4,5,6].…”

A novel hybrid support vector machine with decision tree for data classification

“…The major advantage of decision tree method lies in identifying solutions (Mehenni, 2015). In certain situations when we confront a large sample space, this approach can make data preparation much easier and more understandable for users without technical knowledge compared with remaining methods (Mehenni, 2015).…”