Application of a new FP-network mining algorithm based on association matrix in power system
Fengjie Sun, Chengmin Wang, Weimin Zheng, Fei Chen, Pan Dai, Sen Wang · 2017
Power industry has accumulated a lot of data resources, and the research emphasis is how to improve efficiency of data mining. Aiming at the demerits of FP-tree (frequent pattern tree) mining algorithm, this paper puts forward a FP-network model which innovatively compresses data into a network diagram and it is stored by transaction-item matrix in computers. Node capacity is also innovatively defined to help mining association rules. This method can be applied to power system research and study cases of line faults is established to identify the weak links in grid. Results show that FP-network algorithm not only inherits the merits of FP-growth algorithm, but also only needs to scan the database once. And it is also easy to handle database update and maintenance, thus this method improves the efficiency of association rule mining for big data of power system.