| Summary: | Data mining aims to discover knowledge in large databases. The desired knowledge, normally represented as patterns, are deemed interesting if they benefit some applications. Therefore, the objective of data mining can be translated to finding interesting patterns from observational data. In this thesis, we focus on the simplest form of patterns, which is a set of features (items), also called an itemset. In a high level, data mining process can be split into three parts. The first is to define the notion of interesting patterns. The solution of this subtask is highly application and domain dependent. Having the mathematical formulation of interesting patterns, the next subtask is to find/enumerate those interesting patterns. As the number of interesting patterns are usually too much, the last subtask is to present those patterns in an interpretable and concise form.
|
|---|