Mining Association Rules Inside a Relational Database - A Case Study
Mirela Danubianu, Štefan Gheorghe Pentiuc, Iolanda Tobolcea · International Multi-Conference on Computing in Global Information Technology · 2011
In the context of the necessity to find new knowledge in data, last decade, data mining has become an area of great interest. Although most data mining systems work with data stored in flat files, sometimes it is beneficial to implement data mining algorithms within a DBMS, in order to use SQL or other facilities provided, to discover patterns in data. In this paper we consider a way to discover association rules from data stored into a relational database. We make also a comparative study of performances obtained by applying the following methods: stored procedures in database or candidate and frequent itemsets generated in SQL using a k-way join and a subquery-based algorithm. This study is used to choose the best solution to implement in the particular case of building a dedicated data mining system for personalized therapy of speech disorders optimization.