Distributed random forest method for risk assessment of communication network

The invention discloses a distributed random forest method for risk assessment of a communication network, and the method comprises a data preprocessing stage, an offline training stage of a model andan online prediction stage. The data preprocessing stage comprises the steps: dividing the training data into an optimal number of partitions through the data preprocessing stage, so the proposed model can accelerate parallel and distributed training tasks; dividing the partitioned data into a training set and a test set, then constructing a random forest model for training by utilizing spark, and finally performing online prediction by utilizing the trained model. According to the method, the defects of a conventional communication network risk assessment method are overcome, so the risk assessment result is more reliable; in addition, the big data processing efficiency of a centralized machine learning method is improved, the time cost is saved, the big data processing efficiency in risk assessment is improved,