摘要
异常值检测在数据处理中具有重要意义。为解决电力数据量庞大,维数爆炸问题,文章提出深度自动编码一类支持向量机(AE-OCSVM)模型。该模型使用深度自动编码网络对输入数据降维和进行特征表示,利用OC-SVM对异常值进行预测,采用Isolation Forest、OC-SVM、PCA-KMeans、PCA-GMM(TN=0)、DBSCAN、LOF、DAGMM、VAEGMM和AE-OCSVM 9种算法处理同一组数据,以验证文章所提方法优于其他模型。
Outlier detection is of great significance in data processing.To solve the problem of large data volume and exploding dimensionality,this paper proposes a deep autoencoder support vector machine(AE-OVSVM)model.The model first uses a deep autoencoder network to reduce the dimensionality of the input data for feature representation,and then uses OC-SVM to predict outliers.Finally,9 algorithms including Isolation Forest,OC-SVM,PCA KMeans,PCA-GMM(TN=0),DBSCAN,LOF,DAGMM,VAEGMM,and AE ocsvm were used to process the same set of data,verifying that the proposed method outperforms other models.
出处
《今日自动化》
2024年第11期124-125,128,共3页
Automation Today
