時間序列異常檢測演算法研究?
論文題目是 基於時間序列異常檢測演算法的研究,但是看了很多論文還是無從下手,希望有大神能夠指點一二,感激不盡!
添加:就是在時間序列中查找異常,可以是點異常,也可以是序列異常,這方面的研究是否還可以做,如果可以,我該怎麼做?
取決於所用數據集和技術,關於時間序列異常檢測的資料很多,我給你些做參考吧:
文章/討論
https://blog.twitter.com/2015/introducing-practical-and-robust-anomaly-detection-in-a-time-series
machine learningclustering - Time series and anomaly detectionNetflix』s Surus Anomaly Detection ProjectAnomaly Detection : MachineLearning數據集/工具
https://yahooresearch.tumblr.com/post/114590420346/a-benchmark-dataset-for-time-series-anomalySAX
論文
Assumption-Free Anomaly Detection in Time Serieshttp://alumni.cs.ucr.edu/~ratana/SSDBM05.pdf
An Improved ARIMA-Based Traffic Anomaly Detection Algorithm for Wireless Sensor Networkshttp://www.hindawi.com/journals/ijdsn/2016/9653230/
Anomaly Detection in Real-Valued Multidimensional Time Series
http://www.ase360.org/bitstream/handle/123456789/56/submission34.pdf?sequence=1isAllowed=yNetwork Anomaly Detection Based on Statistical Approach and Time Series Analysis
http://bcmi.sjtu.edu.cn/~huangkai/resource/Network%20Anomaly%20Detection%20Based%20on%20Statistical%20Approach%20and%20Time%20Series%20Analysis.pdfhttps://www.cs.unm.edu/~forrest/publications/noveltydetection96.pdf
Non-parametric anomaly detection
in sentiment time series data
Multi-Resolution Anomaly Detection for the
Internet
http://stat-or.unc.edu/research/Current%20Reports/techpdf/zhang.zhu.et.al.info.pdf
Anomaly Detection in the WIPER System Using Markov Modulated
Poisson Process
Detection of Outliers in Time Series Data
http://epublications.marquette.edu/cgi/viewcontent.cgi?article=1047context=theses_openLocal Anomaly Detection
http://www.jmlr.org/proceedings/papers/v22/saligrama12/saligrama12.pdfDetecting Anomalies in a Time Series Database
https://www.cs.umn.edu/sites/cs.umn.edu/files/tech_reports/09-004.pdfAnomaly detection and prediction of sensors faults in a
refinery using data mining techniques and fuzzy logic
https://umexpert.um.edu.my/file/publication/00005678_85336.pdf
Sensitivity of PCA for Traffic Anomaly Detection
http://ftp.ics.forth.gr/netlab/mobile/Bibliography/LoadBalancing/LB/PCA_Anomaly_Deytection.pdfMultiResolution Anomaly Detection
Method for Long Range Dependent
Time Series
Framework for comparison of
network anomaly detection
algorithms
Symbolic time series analysis via wavelet-based partitioning
http://www.mne.psu.edu/ray/journalAsokRay/2006/171RajagopalanRay06.pdfFinding the Unusual Medical Time Series:
Algorithms and Applications
http://www.cs.cuhk.hk/~adafu/Pub/titb05.pdf
A Framework for Discovering Anomalous Regimes in
Multivariate Time-Series Data with Local Models
Multivariate Online Anomaly Detection Using
Kernel Recursive Least Squares
Combining Filtering and Statistical Methods for Anomaly Detection
https://www.usenix.org/legacy/event/imc05/tech/full_papers/soule/soule.pdfAnomaly Detection in Network Traffic Using
Selected Methods of Time Series Analysis
A Robust Anomaly Detection Technique Using
Combined Statistical Methods
https://hal.archives-ouvertes.fr/hal-00620018/document
Outlier Detection for Temporal Data: A Survey
http://research.microsoft.com/pubs/217054/gupta14_tkde.pdfDetection of abnormal behavior in trade data
using Wavelets, Kalman Filter and Forward Search
Anomaly Detection : A Survey
http://cucis.ece.northwestern.edu/projects/DMS/publications/AnomalyDetection.pdfRANK BASED ANOMALY DETECTION
ALGORITHMS
Mote-based Online Anomaly Detection using
Echo State Networks
http://diku.dk/forskning/Publikationer/tekniske_rapporter/2009/Samlet_teknisk_rapport_09-01.pdf
Anomaly detection from server log
data
A case study
HOT SAX: Finding the Most Unusual Time Series
Subsequence: Algorithms and Applications
源碼
GitHub - twitter/AnomalyDetection: Anomaly Detection with RAnomalyDetectionToolboxGitHub - tadeze/Group-Anomaly-Detection-with-Bayesian-NetworkGitHub - mdlima/MonitoringAnomalyDetection: Anomaly Detection for Yuri"s ProjectGitHub - NanaKwame/AnomalyAlg: Implementation of Anomaly Detection Algorithm in matlab
GitHub - tsaiJN/Proximity_anomaly_detection: this is a matlab implement of paper "Proximity-Based Anomaly Detection using Sparse Structure Learning" Tsuyoshi et al. 2009Downloads – John Quinn教程http://www.autonlab.org/tutorials/biosurv01.pdfhttps://www.siam.org/meetings/sdm13/gupta.pdf可參考下面兩本書:
Adaptive Filtering and Change Detection
http://www.ue.eti.pg.gda.pl/~wrona/lab_dsp/cw07/MATERIALY/Adaptive%20Filtering%20and%20Change%20Detection%20(2000)(Wiley)(Fredrik%20Gustafsson).pdfDetection of Abrupt Change - Theory Applications
ftp://ftp.irisa.fr/local/as/mb/k11.pdf
你的問題很大,情況很多,offline detection還是online detection演算法也不一樣。如果是offline的話,把時間序列就看作是一組sample,統計上有些很簡單的做法,例如基於分位數的做法簡單可靠(Box-and-Whisker Plots: Interquartile Ranges and Outliers),也可參考這篇論文的總結:http://d-scholarship.pitt.edu/7948/1/Seo.pdf。還可以進一步分析一階二階差分里的outlier。去查白聚山(1998,2003)兩篇關於時間序列突變點的經典文獻。當年Eviews還沒這個功能的時候,我用C實現過。
異常檢測之正態分布 這是統計學在異常檢測中的應用
推薦閱讀:
※如何評價facebook開源的prophet時間序列預測工具?
※機器學習門下,有哪些在工業界應用較多,前景較好的小方向?
※時間序列數據的聚類有什麼好方法?
※一個文本里的所有詞不變,順序隨機後,那麼熵改變了多少?如何計算?
※同樣是數據分析方法,為什麼時間序列分析沒有數據挖掘或機器學習那麼火?