Abstract:

The characteristics of power communication network equipment's maintenance,alarm and asset data were analyzed,and then the data preprocessing work was done by adopting attributes reduction and clustering method.The related index and influencing factors required by the analysis of equipment repair work were proposed.Meanwhile,the data mining method which combines the techniques of MapReduce and Apriori algorithms was presented.Finally,the proposed data mining method was applied to mine the actual network data.According to the mined results,the network maintenance work was analyzed from the perspectives of maintenance distribution,maintenance effectiveness on operated service,maintenance time-cost and so forth.

Key words: power telecommunication network, big data technology, maintenance, data preprocessing, data mining