Home
MUSES is the diploma thesis of Thanos Pappas, completed in September 2015 in University of Ioannina Department of Computer Science & Engineering under the supervision of Panos Vassiliadis. MUSES is a tool which receives as input all the changes that the schema of a database has undergone and produces as output frequent patterns from these changes. In order to achieve this goal, sequence mining algorithms that support two different support counting methods were used. In addition, a graphical user interface providing an interactive way to show the frequent patterns to the user was built.
Every software system, including databases, must have the ability to evolve in time in order to adapt to changes resulting new requirements and functionalities. A database is suffering changes on its internal structure during its lifetime: in every transition between 2 different versions of the database, there is a vector of changes (number of attributes inserted, number of attributes deleted etc.). This process is called Schema Evolution and it is the ability of a database system to respond to changes derived from the needs of new applications or the evolution of existing applications. The study of frequent patterns is crucial because it can ease and reduce the cost of maintenance of the database. MUSES is a software that it takes as input all the changes that take place in schema of a database during its lifetime and produce as output frequent patterns, for these changes, applying sequence mining algorithms. But, what is a frequent pattern? A frequent pattern is a sequence of changes that occurs enough number of times satisfying the given constrains.
When we run MUSES we will see the following window:
In order to run the algorithm, we have to load 2 different files. In the first file all the changes of database schema history are stored. We can load the file by clicking: File --> Open
Second file keeps stats about the tables of the database. We can load this file by clicking in the Load button and choose the correct file.
Before executing the algorithm, it is important to set a few parameters such as the counting method:
and minimum support threshold:
Now, everything is set and we are ready to execute the algorithm by clicking in the Run button:
The algorithm produced the output and the frequent patterns are showed in the left list of the window. If we click in one frequent pattern we can see in the middle frame all the tables the frequent pattern appear. Finally, if we click in a table we see in the right table the transitions and the changes that appear in the selected table for the selected pattern.
Finally, we can export the frequent patterns to a csv file by clicking File --> Export to CSV
