Time-series novelty detection using one-class support vector machines

By Junshui Ma, Simon Perkins. Proceedings of ... International Joint Conference on Neural Networks, 2003

411 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Economics and Econometrics, Signal Processing

View Paper: https://www.semanticscholar.org/paper/e50df3d1ab7494a5647fad5d74187fb9d55a06ab

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Charu C. Aggarwal
2. Qiang Yang
3. Vipin Kumar
4. Jindong Wang
5. Philip S. Yu
6. Yue Zhao
7. Yiqiang Chen
8. Jiawei Han

In Economics and Econometrics

1. Arindam Banerjee
2. Jing Gao
3. Yang Yu
4. Jiawei Han
5. Qiang Yang
6. Jongyoun Noh
7. Charu C. Aggarwal
8. Suleyman Serdar Kozat

In Signal Processing

1. Charu C. Aggarwal
2. Vipin Kumar
3. Jiawei Han
4. Yue Zhao
5. Jindong Wang
6. Yiqiang Chen
7. Qiang Yang
8. Manish Gupta

Citations in the corpus

• Sensor-Based Abnormal Human-Activity Detection (2008, methodology): “In general, these approaches can be classified into two categories: similarity-based approaches [1] and model-based approaches [12], [19].”
• PyOD: A Python Toolbox for Scalable Outlier Detection (2019, background): “ulticore LOF (Breunig et al., 2000) Proximity No Yes kNN (Ramaswamy et al., 2000) Proximity No Yes AvgkNN (Angiulli and Pizzuti, 2002) Proximity No Yes CBLOF (He et al., 2003) Proximity Yes No OCSVM (Ma and Perkins, 2003) Linear Model No No LOCI (Papadimitriou et al., 2003) Proximity Yes No PCA (Shy”
• Unsupervised Anomaly Detection With LSTM Neural Networks (2017, methodology): “Thus, their performances significantly depend on the length of this time window so that this approach requires careful selection for the length of the time window to provide a satisfactory performance [8], [10].”
• Learning Memory-Guided Normality for Anomaly Detection (2020, methodology): “, using Markov random field (MRF) [15], a mixture of dynamic textures (MDT) [28], Gaussian regression [4], and one-class classification [39, 27, 14].”
• Anomaly Detection in Time Series: A Comprehensive Evaluation (2022, methodology): “Time series classification, in contrast, describes the process of assigning entire time series to certain classes; it is often used as a post-processing step to classify detected anomalies into domain-specific classes, but some anomaly detection algorithms, such as PS-SVM [85], SR-CNN [112], COPOD [”
• Outlier Detection for Temporal Data: A Survey (2014, methodology): “Clustering Methods: Popular clustering methods include k-Means [25], EM [26], phased k-Means [27], dynamic clustering [24], k-medoids [21], [22], singlelinkage clustering [28], clustering of multi-variate time series in the principal components space [29], oneclass SVM [30], [31], [32], [33] and sel”
• Learning with Augmented Class by Exploiting Unlabeled Data (2014, methodology): “OC-SVM : One-class SVM (Sch¨olkopf et al. 2001) is another state-of-the-art outlier detector (Ma and Perkins 2003), which computes a binary function that is supposed to capture regions in input space where the probability density lives.”
• Unsupervised Deep Anomaly Detection for Multi-Sensor Time-Series Signals (2021, background): “Conventional anomaly detection techniques such as PCA [17], k-means [18], OCSVM [19] and Autoencoder [20] are unable to deal with multivariate timeseries signals since they cannot simultaneously capture the spatial and temporal dependencies.”
• Anomaly detection: A survey (2009, positive): “A transformation technique for time-series data has been proposed by using phase spaces [Ma and Perkins 2003b].”