Outlier Detection Algorithm for Multi-Source Data Based on Improved Density Peak Clustering
The data in the multi-source data set is complex and large,and it is difficult to identify the abnormal data.Aiming at the problem of low accuracy and poor stability of outlier detection in multi-source data,this paper proposes a multi-source data processing method based on the improved peak density clustering algorithm(NDPC algo-rithm)and constructs the NDPC-SVM multi-source data outlier detection model on the basis of this algorithm.First-ly,the model uses data preprocessing to digitize multi-source pose image data to improve the operability of the data;then,it uses a differential privacy protection algorithm to encrypt the data and constructs a privacy data query mecha-nism to improve the privacy of the data;then,it uses the NDPC algorithm to cluster the data and improve the robust-ness of the model.Finally,the NDPC-SVM multi-source data anomaly detection model is constructed by cross-vali-dation optimization.The simulation results of ablation experiments show that the superposition of the four optimization algorithms significantly improves the accuracy and stability of abnormal data detection.Simulation results show that,compared with the baseline clustering algorithm model,the precision of the NDPC-SVM model is as high as 93.14%,the recall is improved by 2.48 on average,and the comprehensive performance is improved by 3.35%.Therefore,the NDPC-SVM multi-source data anomaly detection model based on the NDPC algorithm in this paper not only solves the difficulty of multi-source data processing,but also improves the accuracy and stability of outlier detection.
Density peak clusteringMulti-source dataAnomaly detection
NETL
NSTL
万方数据
