Index-based FP-growth: Distributed data mining

Massive data oriented applications are running on very large databases with the task to extract useful information. Data mining is one of the key applications here. However, most current approaches for mining rules suffer from performance problems and multiple database sources. Particularly their representations are limited for single data source. As a result of this representational limitation, a serious difficulty appears when having large attribute values and single data source. This thesis proposes an index-based representation that resolves these problems. Particularly, we developed a Distributed Association Rule Mining model called “Index-based FP-Growth Method” using the index representation. Experimental results on various datasets show that our index-based association rule mining approach improves the model accuracy and efficiency.