Heterogeneous Distributed Data Mining Method Based on Decision Tree Classification Algorithm

The mining of heterogeneous distributed data is a key technology for insight data traceability and classification feature analysis. A heterogeneous distributed data mining method based on decision tree classification algorithm is proposed, and a structured heterogeneous distributed big data label classification model is constructed. The decision tree shift of heterogeneous distributed data is achieved by combining the label classification statistical method. The decision tree classification algorithm is used to achieve label classification statistics and state transfer of distributed big data. Based on the purity deviation difference and direction deviation, the segmentation values in data label classification are reconstructed and the decision tree structure layer is reorganized. The arithmetic difference of each segmentation block of distributed heterogeneous big data is solved, and more embedding capacity is obtained through histogram shift algorithm. The statistical related parameter feature quantity of data is extracted, and screening is carried out according to feature attributes, using mutual information to implement decision tree classification and big data mining for the feature attributes to be embedded. And simulation tests were conducted on the data mining model under different attack models. The simulation results showed that using this method for heterogeneous distributed data mining can design corresponding shift directions and distances for each heterogeneous data distribution area. There will be no errors in the mining process, which improves the traceability and anti attack ability of the data. Moreover, the distortion of the data is small, and the traceability and classification ability of the data are strong.