Supervised Anomaly Detection in Uncertain Pseudoperiodic Data Streams

Ma, Jiangang; Sun, Le; Wang, Hua; Zhang, Yanchun; Aickelin, Uwe

doi:10.1145/2806890

Cited by 85 publications

(53 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the algorithm performance is not resilient as different parameters selections will have a huge influence on the detection results. [31] introduced a new scheme to detect outliers in noisy data streams by employing a wavelet based soft-threshold filtering approach that can remove uncertainties in time series data streams, However, the method cannot always ensure high accuracy on a variety of real datasets. Recently, the work of [32] applied outlier detection in mobile computing via machine learning based clustering techniques.…”

Section: Related Workmentioning

confidence: 99%

LSTM Learning With Bayesian and Gaussian Processing for Anomaly Detection in Industrial IoT

Jiang

Xie

et al. 2020

IEEE Trans. Ind. Inf.

209

View full text Add to dashboard Cite

The data generated by millions of sensors in Industrial Internet of Things (IIoT) is extremely dynamic, heterogeneous, and large scale. It poses great challenges on the real-time analysis and decision making for anomaly detection in IIoT. In this paper, we propose a LSTM-Gauss-NBayes method, which is a synergy of the long short-term memory neural network (LSTM-NN) and the Gaussian Bayes model for outlier detection in IIoT. In a nutshell, the LSTM-NN builds model on normal time series. It detects outliers by utilising the predictive error for the Gaussian Naive Bayes model. Our method exploits advantages of both LSTM and Gaussian Naive Bayes models, which not only has strong prediction capability of LSTM for future time point data, but also achieves an excellent classification performance of Gaussian Naive Bayes model through the predictive error. We evaluate our approaches on 3 real-life datasets that involve both long-term and short-term time-dependency. Empirical studies demonstrate that our proposed techniques outperform the bestknown competitors, which is a preferable choice for detecting anomalies.

show abstract

Section: Related Workmentioning

confidence: 99%

LSTM Learning With Bayesian and Gaussian Processing for Anomaly Detection in Industrial IoT

Jiang

Xie

et al. 2020

IEEE Trans. Ind. Inf.

209

View full text Add to dashboard Cite

show abstract

“…1 Among the many PhysioBank datasets, we chose the 102nd record of the MIT-BiH Arrhyhmia Database (mitdb). The mitdb is widely used for testing anomaly detection [3,10,20]. According to an annotation to this dataset, a pacemaker is used most of the time, but from time 29,276 to time 62,531 the dataset corresponds to regular heartbeats.…”

Section: Ecgmentioning

confidence: 99%

dLSTM: a new approach for anomaly detection using deep learning with delayed prediction

Maya

Ueno

Nishikawa

2019

Int J Data Sci Anal

View full text Add to dashboard Cite

In this paper, we propose delayed Long Short-Term Memory (dLSTM), an anomaly detection method for time-series data. We first build a predictive model from normal (non-anomalous) training data, then perform anomaly detection based on the prediction error for observed data. However, there are multiple states in the waveforms of normal data, which may lower prediction accuracy. To deal with this problem, we utilize multiple prediction models based on LSTM for anomaly detection. In this scheme, the prediction accuracy strongly depends on the method of selecting a proper predictive model from multiple possible models. We propose a novel method to determine the proper predictive model for anomaly detection. Our approach provides multiple predicted value candidates in advance and selects the one that is closest to the measured value. We delay the model selection until the corresponding measured values are acquired. Using this concept for anomaly detection, dLSTM selects the proper predictive model to enhance prediction accuracy. In our experimental evaluation using real and artificial data, dLSTM detects anomalies more accurately than methods in comparison.

show abstract

“…Note that, on one hand, similar to other real-world data streams, such as RFID data streams [9], and electrocardiogram streams [41], microblog streams are also full of large amount of noise. on one hand, different from news documents,That is to say, microblog content does not always relate to a social, political or sport event.…”

Section: High Quality Microblog Streammentioning

confidence: 99%

A probabilistic method for emerging topic tracking in Microblog stream

et al. 2016

Self Cite

View full text Add to dashboard Cite

Microblog is a popular and open platform for discovering and sharing the latest news about social issues and daily life. The quickly-updated microblog streams make it urgent to develop an effective tool to monitor such streams. Emerging topic tracking is one of such tools to reveal which new events are attracting the most online attention at present. However, due to the fast changing, high noise and short length of microblog feeds, two challenges should be addressed in emerging topic tracking. One is the problem of detecting emerging topics early, long before they become hot, and the other is how to effectively monitor evolving topics over time. In this study, we propose a novel emerging topics tracking method, which aligns emerging word detection from temporal perspective with coherent topic mining from spatial perspective. Specifically, we first design metrics to estimate word novelty and fading based on local weighted linear regression (LWLR), which can highlight emerging-topic-relevant words and suppress existing-topic-relevant words. We then track emerging topics by leveraging topic novelty and fading probabilities, which are learnt by designing and solving an optimization problem. We evaluate our method on a microblog stream containing over one million feeds. Experimental results show the promising performance of the proposed method in detecting emerging topic and tracking topic evolution over time on both effectiveness and efficiency.

show abstract

Supervised Anomaly Detection in Uncertain Pseudoperiodic Data Streams

Cited by 85 publications

References 45 publications

LSTM Learning With Bayesian and Gaussian Processing for Anomaly Detection in Industrial IoT

LSTM Learning With Bayesian and Gaussian Processing for Anomaly Detection in Industrial IoT

dLSTM: a new approach for anomaly detection using deep learning with delayed prediction

A probabilistic method for emerging topic tracking in Microblog stream

Contact Info

Product

Resources

About