Application of C5.0 algorithm in failure prediction of smart meters
Jincheng Yang, Jiang Ping, Chen Guangyu, Yuan Tiejiang, Fei Xue · 2016
Aiming at the Smart meters failure prediction problem and based on historical failure data of smart meters in a region of Xinjiang, a smart meter fault identification model is proposed based on C5.0 algorithm: first, after data preprocessing of smart meters history failure database is divided into two parts, training set and testing set; secondly, using C5.0 algorithm for data mining of training set, spanning the smart meters failure decision tree and preliminary prediction rules; additionally, the correctness of the preliminary prediction rules is evaluated by the data of the test set, and if the accuracy meets the requirement, the prediction rules are determined and if not, then return to the training set again; finally, smart meters failure prediction model generated by the definitive prediction rules. The results of examples showed that accuracy of failure prediction model for the smart meters is higher, achieving good prediction effect, and this method can provide some reference for further research on failure prediction of smart meters.