Research on abnormal detection of one-class support vector machine based on ensemble cooperative semi-supervised learning

Abstract

In order to promote the accuracy of anomaly detection model under the condition of only a small number of labeled samples and large number of unlabeled samples, abnormal detection of One-class Support Vector Machine(SVM) based on ensemble cooperative Semi-supervised Learning is proposed. A kind of One-class SVM model which bring supervision with a small number of abnormal samples can classify samples with max interval. The semi-supervised learning methods easily suffer from the low accuracy because the mistake labeled sample are chosen as training sample set. Refer to the semi-supervision method of Tri-training, the K-Nearest Neighbour(KNN) and Naive Bayes classifier are used to uses to assist the One-class SVM based on ensemble cooperative Semi-supervised learning method which can classify the large number of unlabeled samples as accurate as possible. The weight is also given after ensemble cooperative Semi-supervised Learning. Then the proposed semi-supervised One-class SVM would be trained with the result and used to classify test samples. The experimental results on UCI dataset show that the proposed algorithm achieves higher classification accuracy with less labeled samples and it improves generalization performance and reduces the labelling cost.

A Deep-Convolutional-Neural-Network-Based Semi-Supervised Learning Method for Anomaly Crack Detection

“…Semi-supervised learning combines supervised learning with unsupervised learning together, and it can train classifiers with few labeled samples. Typical semi-supervised learning algorithms, for example, self-training [21], hybrid models [22], graph-based ones [23] and SVM-based [24] ones can be applied to wall crack detection [25], pavement crack detection [26] and steel structure surface defect detection [27]. During the training, classifiers are trained by a small amount of labeled data and then are employed to classify a great amount of unlabeled data.…”