To improve the accuracy and anti-interference ability for risk analysis of mobile device

and reduce the complaint rate

this paper proposed a model of risk analysis base on PCA and BP neural network.The feature vectors of device risk types are reduces from 10 to 3 and main information of raw data can be got by PCA.And then built a device risk analysis model using BP neural network which the input data is the low dimension feature vectors got from PCA.Experiments compared performances of our model and some other models

the results show the model combine PCA and BP neural network can identify risks accurately.