Data Mining Fault Information Reports for Prediction

The paper discusses an application of datamining techniques for analysis of fault information data of transmission network in New Zealand. The analysis is mainly aimed to provide necessary prediction knowledge of a fault after an event. This prediction could be a reason for fault, faulted equipment, estimated loss of supply and estimated time for restoration. Our data used for this paper contains transmission faults of approximately 342 events spanning over three years. In this paper, we introduce the general datamining procedure and discuss its application to analyse the fault information reports. Our solution methodology, pre-processing data selection, analysis algorithms are also discussed. Finally we present and analyse some of the results