欢迎来到 PyOD 文档！¶

部署 & 文档 & 统计 & 许可证

先阅读我¶

欢迎使用 PyOD，这是一个全面但易于使用的 Python 库，用于检测多元数据中的异常。无论您处理的是小规模项目还是大型数据集，PyOD 都提供了一系列算法来满足您的需求。

对于时间序列的异常检测，请使用 TODS。
对于图异常检测，请使用 PyGOD。
性能比较与数据集: 我们有一份45页的全面异常检测基准论文。完全开源的ADBench 在57个基准数据集上比较了30种异常检测算法。
了解更多关于异常检测 请访问异常检测资源
分布式系统中的PyOD: 你也可以在 Databricks上运行PyOD。

关于 PyOD¶

PyOD，成立于2017年，已成为用于**检测多元数据中的异常/离群对象**的**Python库**。这个既令人兴奋又具有挑战性的领域通常被称为离群检测或异常检测。

PyOD 包含超过 50 种检测算法，从经典的 LOF (SIGMOD 2000) 到前沿的 ECOD 和 DIF (TKDE 2022 和 2023)。自 2017 年以来，PyOD 已成功应用于众多学术研究项目和商业产品，下载量超过 2200 万次。它也得到了机器学习社区的广泛认可，拥有各种专门的帖子/教程，包括 Analytics Vidhya、KDnuggets 和 Towards Data Science。

PyOD 的特点包括：

统一、用户友好的界面 跨多种算法。
广泛的模型，从经典技术到 PyTorch 中的最新深度学习方法。
高性能与效率，利用 numba 和 joblib 进行即时编译和并行处理。
快速训练与预测，通过 SUOD 框架实现 [AZHC+21]。

用5行代码进行异常检测:

# Example: Training an ECOD detector
from pyod.models.ecod import ECOD
clf = ECOD()
clf.fit(X_train)
y_train_scores = clf.decision_scores_  # Outlier scores for training data
y_test_scores = clf.decision_function(X_test)  # Outlier scores for test data

选择合适的算法： 不确定从哪里开始？考虑这些稳健且可解释的选项：

ECOD: 使用ECOD进行异常检测的示例
Isolation Forest: 使用 Isolation Forest 进行异常检测的示例

或者，探索 MetaOD 以采用数据驱动的方法。

引用 PyOD:

PyOD 论文发表在机器学习研究杂志 (JMLR) (MLOSS 专栏)。如果您在科学出版物中使用 PyOD，我们将非常感谢您引用以下论文:

@article{zhao2019pyod,
    author  = {Zhao, Yue and Nasrullah, Zain and Li, Zheng},
    title   = {PyOD: A Python Toolbox for Scalable Outlier Detection},
    journal = {Journal of Machine Learning Research},
    year    = {2019},
    volume  = {20},
    number  = {96},
    pages   = {1-7},
    url     = {http://jmlr.org/papers/v20/19-011.html}
}

或者:

Zhao, Y., Nasrullah, Z. and Li, Z., 2019. PyOD: A Python Toolbox for Scalable Outlier Detection. Journal of machine learning research (JMLR), 20(96), pp.1-7.

关于异常检测的更广泛视角，请参阅我们的 NeurIPS 论文 ADBench: 异常检测基准论文和 ADGym: 深度异常检测的设计选择:

@article{han2022adbench,
    title={Adbench: Anomaly detection benchmark},
    author={Han, Songqiao and Hu, Xiyang and Huang, Hailiang and Jiang, Minqi and Zhao, Yue},
    journal={Advances in Neural Information Processing Systems},
    volume={35},
    pages={32142--32159},
    year={2022}
}

@article{jiang2023adgym,
    title={ADGym: Design Choices for Deep Anomaly Detection},
    author={Jiang, Minqi and Hou, Chaochuan and Zheng, Ao and Han, Songqiao and Huang, Hailiang and Wen, Qingsong and Hu, Xiyang and Zhao, Yue},
    journal={Advances in Neural Information Processing Systems},
    volume={36},
    year={2023}
}

ADBench 基准测试和数据集¶

我们刚刚发布了一份45页的，最全面的 ADBench: Anomaly Detection Benchmark [#Han2022ADBench]_。完全开源的ADBench 在57个基准数据集上比较了30种异常检测算法。

ADBench 的组织结构如下：

为了更简单的可视化，我们通过 compare_all_models.py进行了**选定模型的比较**。

已实现的算法¶

PyOD 工具包由三个主要功能组组成：

(i) 个体检测算法 :

类型	缩写	算法	年	类	参考
概率性	ECOD	使用经验累积分布函数进行无监督异常检测	2022	`pyod.models.ecod.ECOD`	[ALZH+22]
概率性	COPOD	COPOD: 基于Copula的异常检测	2020	`pyod.models.copod.COPOD`	[ALZB+20]
概率性	ABOD	基于角度的异常检测	2008	`pyod.models.abod.ABOD`	[AKZ+08]
概率性	FastABOD	基于角度的异常检测使用近似快速检测	2008	`pyod.models.abod.ABOD`	[AKZ+08]
概率性	MAD	中位数绝对偏差 (MAD)	1993	`pyod.models.mad.MAD`	[AIH93]
概率性	SOS	随机异常选择	2012	`pyod.models.sos.SOS`	[AJHuszarPvdH12]
概率性	QMCD	准蒙特卡罗离差异常检测	2001	`pyod.models.qmcd.QMCD`	[AFM01]
概率性	KDE	使用核密度函数进行异常检测	2007	`pyod.models.kde.KDE`	[ALLP07]
概率性	采样	通过采样进行基于距离的快速异常检测	2013	`pyod.models.sampling.Sampling`	[ASB13]
概率性	GMM	概率混合模型用于异常分析		`pyod.models.gmm.GMM`	[AAgg15] [Ch.2]
线性模型	PCA	主成分分析（加权投影距离到特征向量超平面的总和）	2003	`pyod.models.pca.PCA`	[ASCSC03]
线性模型	KPCA	核主成分分析	2007	`pyod.models.kpca.KPCA`	[AHof07]
线性模型	MCD	最小协方差行列式（使用马氏距离作为异常值分数）	1999	`pyod.models.mcd.MCD`	[AHR04, ARD99]
线性模型	CD	使用 Cook’s 距离进行异常值检测	1977	`pyod.models.cd.CD`	[ACoo77]
线性模型	OCSVM	一类支持向量机	2001	`pyod.models.ocsvm.OCSVM`	[AScholkopfPST+01]
线性模型	LMDD	基于偏差的异常检测 (LMDD)	1996	`pyod.models.lmdd.LMDD`	[AAAR96]
基于接近性	LOF	局部离群因子	2000	`pyod.models.lof.LOF`	[ABKNS00]
基于接近性	COF	基于连通性的离群因子	2002	`pyod.models.cof.COF`	[ATCFC02]
基于接近性	Incr. COF	内存高效连通性基异常因子（速度较慢但降低存储复杂度）	2002	`pyod.models.cof.COF`	[ATCFC02]
基于接近性	CBLOF	基于聚类的局部异常因子	2003	`pyod.models.cblof.CBLOF`	[AHXD03]
基于接近性	LOCI	LOCI：使用局部相关积分进行快速异常检测	2003	`pyod.models.loci.LOCI`	[APKGF03]
基于接近性	HBOS	基于直方图的异常分数	2012	`pyod.models.hbos.HBOS`	[AGD12]
基于接近性	kNN	k 近邻 (使用到第 k 个最近邻的距离作为异常分数)	2000	`pyod.models.knn.KNN`	[AAP02, ARRS00]
基于接近性	AvgKNN	平均 kNN（使用到 k 个最近邻居的平均距离作为异常分数）	2002	`pyod.models.knn.KNN`	[AAP02, ARRS00]
基于接近性	MedKNN	中位数 kNN (使用到 k 个最近邻的中位数距离作为异常分数)	2002	`pyod.models.knn.KNN`	[AAP02, ARRS00]
基于接近性	SOD	子空间异常检测	2009	`pyod.models.sod.SOD`	[AKKrogerSZ09]
基于接近性	ROD	基于旋转的异常检测	2020	`pyod.models.rod.ROD`	[AABC20]
离群点集成	IForest	隔离森林	2008	`pyod.models.iforest.IForest`	[ALTZ08, ALTZ12]
离群点集成	`INNE`	基于隔离的异常检测使用最近邻集成	2018	`pyod.models.inne.INNE`	[ABTA+18]
离群点集成	DIF	深度隔离森林用于异常检测	2023	`pyod.models.dif.DIF`	[]
离群点集成	FB	特征装袋	2005	`pyod.models.feature_bagging.FeatureBagging`	[]
离群点集成	LSCP	LSCP: 局部选择性组合并行异常集成	2019	`pyod.models.lscp.LSCP`	[AZNHL19]
离群点集成	XGBOD	基于极端提升的异常检测 (有监督)	2018	`pyod.models.xgbod.XGBOD`	[AZH18]
离群点集成	LODA	轻量级在线异常检测器	2016	`pyod.models.loda.LODA`	[APevny16]
离群点集成	SUOD	SUOD: 加速大规模无监督异构异常检测 (加速)	2021	`pyod.models.suod.SUOD`	[AZHC+21]
神经网络	自动编码器	全连接自编码器（使用重构误差作为异常分数）	2015	`pyod.models.auto_encoder.AutoEncoder`	[AAgg15] [Ch.3]
神经网络	VAE	变分自编码器（使用重构误差作为异常分数）	2013	`pyod.models.vae.VAE`	[AKW13]
神经网络	Beta-VAE	变分自编码器（通过调整gamma和容量来定制所有损失项）	2018	`pyod.models.vae.VAE`	[ABHP+18]
神经网络	SO_GAAL	单目标生成对抗主动学习	2019	`pyod.models.so_gaal.SO_GAAL`	[ALLZ+19]
神经网络	MO_GAAL	多目标生成对抗主动学习	2019	`pyod.models.mo_gaal.MO_GAAL`	[ALLZ+19]
神经网络	DeepSVDD	深度一类分类	2018	`pyod.models.deep_svdd.DeepSVDD`	[ARVG+18]
神经网络	AnoGAN	使用生成对抗网络进行异常检测	2017	`pyod.models.anogan.AnoGAN`	[ASSeebockW+17]
神经网络	ALAD	对抗学习异常检测	2018	`pyod.models.alad.ALAD`	[AZRF+18]
神经网络	DevNet	使用偏差网络进行深度异常检测	2019	`pyod.models.devnet.DevNet`	[APSVDH19]
神经网络	AE1SVM	基于自编码器的一类支持向量机	2019	`pyod.models.ae1svm.AE1SVM`	[ANV19]
基于图形的	R-图	通过 R-graph 进行异常检测	2017	`pyod.models.rgraph.RGraph`	[AYRV17]
基于图形的	LUNAR	LUNAR: 通过图神经网络统一局部异常检测方法	2022	`pyod.models.lunar.LUNAR`	[AGHNN22]

(ii) 异常值集合 & 异常值检测器组合框架:

类型	缩写	算法	年	参考
离群点集成		特征装袋	2005	`pyod.models.feature_bagging.FeatureBagging`	[]
离群点集成	LSCP	LSCP: 局部选择性组合并行异常集成	2019	`pyod.models.lscp.LSCP`	[AZNHL19]
离群点集成	XGBOD	基于极端提升的异常检测 (有监督)	2018	`pyod.models.xgbod.XGBOD`	[AZH18]
离群点集成	LODA	轻量级在线异常检测器	2016	`pyod.models.loda.LODA`	[APevny16]
离群点集成	SUOD	SUOD: 加速大规模无监督异构异常检测 (加速)	2021	`pyod.models.suod.SUOD`	[AZHC+21]
组合	平均	通过平均分数进行简单组合	2015	`pyod.models.combination.average()`	[AAS15]
组合	加权平均	通过平均得分与检测器权重进行简单组合	2015	`pyod.models.combination.average()`	[AAS15]
组合	最大化	通过取最大分数进行简单组合	2015	`pyod.models.combination.maximization()`	[AAS15]
组合	AOM	最大值的平均值	2015	`pyod.models.combination.aom()`	[AAS15]
组合	MOA	最大平均值	2015	`pyod.models.combination.moa()`	[AAS15]
组合	中位数	通过取分数的中位数进行简单组合	2015	`pyod.models.combination.median()`	[AAS15]
组合	多数投票	通过取标签的多数投票进行简单组合（可以使用权重）	2015	`pyod.models.combination.majority_vote()`	[AAS15]

(iii) 实用函数:

类型	名称	函数
数据	`pyod.utils.data.generate_data()`	合成数据生成；正常数据由多元高斯分布生成，异常数据由均匀分布生成
数据	`pyod.utils.data.generate_data_clusters()`	集群中的合成数据生成；通过多个集群可以创建更复杂的数据模式
统计	`pyod.utils.stat_models.wpearsonr()`	计算两个样本的加权皮尔逊相关系数
工具	`pyod.utils.utility.get_label_n()`	通过将前 n 个异常分数赋值为 1，将原始异常分数转换为二进制标签
工具	`pyod.utils.utility.precision_n_scores()`	计算精度 @ 排名 n

API 速查表与参考¶

以下API适用于所有检测器模型，便于使用。

pyod.models.base.BaseDetector.fit(): 拟合检测器。在无监督方法中，y 被忽略。
pyod.models.base.BaseDetector.decision_function(): 使用拟合的检测器预测X的原始异常分数。
pyod.models.base.BaseDetector.predict(): 使用拟合的检测器预测特定样本是否为异常值。
pyod.models.base.BaseDetector.predict_proba(): 使用拟合的检测器预测样本为异常值的概率。
pyod.models.base.BaseDetector.predict_confidence(): 预测模型的样本置信度（在 predict 和 predict_proba 中可用）。

拟合模型的关键属性：

pyod.models.base.BaseDetector.decision_scores_: 训练数据的外部得分。得分越高，越异常。异常值往往有更高的得分。
pyod.models.base.BaseDetector.labels_: 训练数据的二进制标签。0 表示内点，1 表示异常点/异常。

参考文献

[AAgg15] (1,2)

Charu C Aggarwal. Outlier analysis. In Data mining, 75–79. Springer, 2015.

[AAS15] (1,2,3,4,5,6,7)

Charu C Aggarwal and Saket Sathe. Theoretical foundations and algorithms for outlier ensembles. ACM SIGKDD Explorations Newsletter, 17(1):24–47, 2015.

[AABC20]

Yahya Almardeny, Noureddine Boujnah, and Frances Cleary. A novel outlier detection method for multivariate data. IEEE Transactions on Knowledge and Data Engineering, 2020.

[AAP02] (1,2,3)

Fabrizio Angiulli and Clara Pizzuti. Fast outlier detection in high dimensional spaces. In European Conference on Principles of Data Mining and Knowledge Discovery, 15–27. Springer, 2002.

[AAAR96]

Andreas Arning, Rakesh Agrawal, and Prabhakar Raghavan. A linear method for deviation detection in large databases. In KDD, volume 1141, 972–981. 1996.

[ABTA+18]

Tharindu R Bandaragoda, Kai Ming Ting, David Albrecht, Fei Tony Liu, Ye Zhu, and Jonathan R Wells. Isolation-based anomaly detection using nearest-neighbor ensembles. Computational Intelligence, 34(4):968–998, 2018.

[ABKNS00]

Markus M Breunig, Hans-Peter Kriegel, Raymond T Ng, and Jörg Sander. Lof: identifying density-based local outliers. In ACM sigmod record, volume 29, 93–104. ACM, 2000.

[ABHP+18]

Christopher P Burgess, Irina Higgins, Arka Pal, Loic Matthey, Nick Watters, Guillaume Desjardins, and Alexander Lerchner. Understanding disentangling in betvae. arXiv preprint arXiv:1804.03599, 2018.

[ACoo77]

R Dennis Cook. Detection of influential observation in linear regression. Technometrics, 19(1):15–18, 1977.

[AFM01]

Kai-Tai Fang and Chang-Xing Ma. Wrap-around l2-discrepancy of random sampling, latin hypercube and uniform designs. Journal of complexity, 17(4):608–624, 2001.

[AGD12]

Markus Goldstein and Andreas Dengel. Histogram-based outlier score (hbos): a fast unsupervised anomaly detection algorithm. KI-2012: Poster and Demo Track, pages 59–63, 2012.

[AGHNN22]

Adam Goodge, Bryan Hooi, See-Kiong Ng, and Wee Siong Ng. Lunar: unifying local outlier detection methods via graph neural networks. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, 6737–6745. 2022.

[AHR04]

Johanna Hardin and David M Rocke. Outlier detection in the multiple cluster setting using the minimum covariance determinant estimator. Computational Statistics & Data Analysis, 44(4):625–638, 2004.

[AHXD03]

Zengyou He, Xiaofei Xu, and Shengchun Deng. Discovering cluster-based local outliers. Pattern Recognition Letters, 24(9-10):1641–1650, 2003.

[AHof07]

Heiko Hoffmann. Kernel pca for novelty detection. Pattern recognition, 40(3):863–874, 2007.

[AIH93]

Boris Iglewicz and David Caster Hoaglin. How to detect and handle outliers. Volume 16. Asq Press, 1993.

[AJHuszarPvdH12]

JHM Janssens, Ferenc Huszár, EO Postma, and HJ van den Herik. Stochastic outlier selection. Technical Report, Technical report TiCC TR 2012-001, Tilburg University, Tilburg Center for Cognition and Communication, Tilburg, The Netherlands, 2012.

[AKW13]

Diederik P Kingma and Max Welling. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013.

[AKKrogerSZ09]

Hans-Peter Kriegel, Peer Kröger, Erich Schubert, and Arthur Zimek. Outlier detection in axis-parallel subspaces of high dimensional data. In Pacific-Asia Conference on Knowledge Discovery and Data Mining, 831–838. Springer, 2009.

[AKZ+08] (1,2)

Hans-Peter Kriegel, Arthur Zimek, and others. Angle-based outlier detection in high-dimensional data. In Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining, 444–452. ACM, 2008.

[ALLP07]

Longin Jan Latecki, Aleksandar Lazarevic, and Dragoljub Pokrajac. Outlier detection with kernel density functions. In International Workshop on Machine Learning and Data Mining in Pattern Recognition, 61–75. Springer, 2007.

[ALK05] (1,2)

Aleksandar Lazarevic and Vipin Kumar. Feature bagging for outlier detection. In Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining, 157–166. ACM, 2005.

[ALZB+20]

Zheng Li, Yue Zhao, Nicola Botta, Cezar Ionescu, and Xiyang Hu. COPOD: copula-based outlier detection. In IEEE International Conference on Data Mining (ICDM). IEEE, 2020.

[ALZH+22]

Zheng Li, Yue Zhao, Xiyang Hu, Nicola Botta, Cezar Ionescu, and H. George Chen. Ecod: unsupervised outlier detection using empirical cumulative distribution functions. IEEE Transactions on Knowledge and Data Engineering, 2022.

[ALTZ08]

Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation forest. In Data Mining, 2008. ICDM'08. Eighth IEEE International Conference on, 413–422. IEEE, 2008.

[ALTZ12]

Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation-based anomaly detection. ACM Transactions on Knowledge Discovery from Data (TKDD), 6(1):3, 2012.

[ALLZ+19] (1,2)

Yezheng Liu, Zhe Li, Chong Zhou, Yuanchun Jiang, Jianshan Sun, Meng Wang, and Xiangnan He. Generative adversarial active learning for unsupervised outlier detection. IEEE Transactions on Knowledge and Data Engineering, 2019.

[ANV19]

Minh-Nghia Nguyen and Ngo Anh Vien. Scalable and interpretable one-class svms with deep learning and random fourier features. In Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2018, Dublin, Ireland, September 10–14, 2018, Proceedings, Part I 18, 157–172. Springer, 2019.

[APSVDH19]

Guansong Pang, Chunhua Shen, and Anton Van Den Hengel. Deep anomaly detection with deviation networks. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 353–362. 2019.

[APKGF03]

Spiros Papadimitriou, Hiroyuki Kitagawa, Phillip B Gibbons, and Christos Faloutsos. Loci: fast outlier detection using the local correlation integral. In Data Engineering, 2003. Proceedings. 19th International Conference on, 315–326. IEEE, 2003.

[APVD20]

Lorenzo Perini, Vincent Vercruyssen, and Jesse Davis. Quantifying the confidence of anomaly detectors in their example-wise predictions. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 227–243. Springer, 2020.

[APevny16] (1,2)

Tomáš Pevn\`y. Loda: lightweight on-line detector of anomalies. Machine Learning, 102(2):275–304, 2016.

[ARRS00] (1,2,3)

Sridhar Ramaswamy, Rajeev Rastogi, and Kyuseok Shim. Efficient algorithms for mining outliers from large data sets. In ACM Sigmod Record, volume 29, 427–438. ACM, 2000.

[ARD99]

Peter J Rousseeuw and Katrien Van Driessen. A fast algorithm for the minimum covariance determinant estimator. Technometrics, 41(3):212–223, 1999.

[ARVG+18]

Lukas Ruff, Robert Vandermeulen, Nico Görnitz, Lucas Deecke, Shoaib Siddiqui, Alexander Binder, Emmanuel Müller, and Marius Kloft. Deep one-class classification. International conference on machine learning, 2018.

[ASSeebockW+17]

Thomas Schlegl, Philipp Seeböck, Sebastian M Waldstein, Ursula Schmidt-Erfurth, and Georg Langs. Unsupervised anomaly detection with generative adversarial networks to guide marker discovery. In International conference on information processing in medical imaging, 146–157. Springer, 2017.

[AScholkopfPST+01]

Bernhard Schölkopf, John C Platt, John Shawe-Taylor, Alex J Smola, and Robert C Williamson. Estimating the support of a high-dimensional distribution. Neural computation, 13(7):1443–1471, 2001.

[ASCSC03]

Mei-Ling Shyu, Shu-Ching Chen, Kanoksri Sarinnapakorn, and LiWu Chang. A novel anomaly detection scheme based on principal component classifier. Technical Report, MIAMI UNIV CORAL GABLES FL DEPT OF ELECTRICAL AND COMPUTER ENGINEERING, 2003.

[ASB13]

Mahito Sugiyama and Karsten Borgwardt. Rapid distance-based outlier detection via sampling. Advances in neural information processing systems, 2013.

[ATCFC02] (1,2)

Jian Tang, Zhixiang Chen, Ada Wai-Chee Fu, and David W Cheung. Enhancing effectiveness of outlier detections for low density patterns. In Pacific-Asia Conference on Knowledge Discovery and Data Mining, 535–548. Springer, 2002.

[AYRV17]

Chong You, Daniel P Robinson, and René Vidal. Provable self-representation based outlier detection in a union of subspaces. In Proceedings of the IEEE conference on computer vision and pattern recognition, 3395–3404. 2017.

[AZRF+18]

Houssam Zenati, Manon Romain, Chuan-Sheng Foo, Bruno Lecouat, and Vijay Chandrasekhar. Adversarially learned anomaly detection. In 2018 IEEE International conference on data mining (ICDM), 727–736. IEEE, 2018.

[AZH18] (1,2)

Yue Zhao and Maciej K Hryniewicki. Xgbod: improving supervised outlier detection with unsupervised representation learning. In International Joint Conference on Neural Networks (IJCNN). IEEE, 2018.

[AZHC+21] (1,2,3,4,5)

Yue Zhao, Xiyang Hu, Cheng Cheng, Cong Wang, Changlin Wan, Wen Wang, Jianing Yang, Haoping Bai, Zheng Li, Cao Xiao, Yunlong Wang, Zhi Qiao, Jimeng Sun, and Leman Akoglu. Suod: accelerating large-scale unsupervised heterogeneous outlier detection. Proceedings of Machine Learning and Systems, 2021.

[AZNHL19] (1,2)

Yue Zhao, Zain Nasrullah, Maciej K Hryniewicki, and Zheng Li. LSCP: locally selective combination in parallel outlier ensembles. In Proceedings of the 2019 SIAM International Conference on Data Mining, SDM 2019, 585–593. Calgary, Canada, May 2019. SIAM. URL: https://doi.org/10.1137/1.9781611975673.66, doi:10.1137/1.9781611975673.66.