Fast Outlier Detection in High Dimensional Spaces

By Fabrizio Angiulli, Clara Pizzuti. European Conference on Principles of Data Mining and Knowledge Discovery, 2002

946 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Biomedical Engineering, Civil and Structural Engineering

View Paper: https://www.semanticscholar.org/paper/290ff9a05058ec14e4cfa7bdb716f5b007ee06eb

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

In Biomedical Engineering

In Civil and Structural Engineering

Citations in the corpus

Mining distance-based outliers in near linear time with randomization and a simple pruning rule (2003, methodology): “[2] F. Angiulli and C. Pizzuti.”
Progress in Outlier Detection Techniques: A Survey (2019, background): “This study was motivated by the fact that most OD methods focus on identifying global outliers, and most of these methods [119], [120] are very sensitive to parameter changes.”
Graph Neural Network-Based Anomaly Detection in Multivariate Time Series (2021, methodology): “• KNN: K Nearest Neighbors uses each point’s distance to its kth nearest neighbor as an anomaly score (Angiulli and Pizzuti 2002).”
HRN: A Holistic Approach to One Class Learning (2020, neutral): “…can be classiﬁed into the following categories: distance based methods [42, 3, 28, 31, 60], density based methods [38, 56, 6], mixture models [3, 43], one-class classiﬁcation based methods [75, 78, 39], deep learning based representation learning using auto-encoders [10, 89, 69, 7, 94] and…”
Deep Learning for Anomaly Detection (2020, background): “blication date: July 202x. 16 Pang, et al. effective distance-based anomaly measures introduced,e.g., DB outliers [73, 74], k-nearest neighbor distance [126, 127], average k-nearest neighbor distance [9], relative distance [174], and random nearest neighbor distance [117, 145]. One major limitation ”
Enhancing one-class support vector machines for unsupervised anomaly detection (2013, methodology): “Here, global methods, for example the k-nearest neighbor [23, 2] and local methods exist.”
Deep Learning for Anomaly Detection in Time-Series Data: Review, Analysis, and Guidelines (2021, methodology): “Classical methods including linear model-based methods [3], distance-based methods [4], density-based methods [5], and support vector machines [6], are still a viable choice of algorithm.”
Learning Graph Structures With Transformer for Multivariate Time-Series Anomaly Detection in IoT (2021, methodology): “3) Baselines: We compare our GTA with a wide range of state-of-the-arts in multivariate time series anomaly detection, including: (1) reconstruction-based models: PCA, AE [10], KitNet [57], DAGMM [12], GAN-Li [20], OmniAnomaly [13], LSTM-VAE [11], MAD-GAN [6], and (2) forecasting-9 based models: KNN”
A Comparative Evaluation of Unsupervised Anomaly Detection Algorithms for Multivariate Data (2016, methodology): “Then, an anomaly score is computed using these neighbors, whereas two possibilities have been proposed: Either the distance to the k-nearest-neighbor is used (a single one) [41] or the average distance to all of the k-nearest-neighbors [42] is computed.”
Outlier mining in large high-dimensional data sets (2005, neutral): “✦”
ADBench: Anomaly Detection Benchmark (2022, background): “The presence of duplicated anomalies is also called the “anomaly masking” [55, 60, 100], posing challenges to many AD algorithms [25], e.g., the density-based KNN [11, 144].”
Interpreting and Unifying Outlier Scores (2011, neutral): “Scoring variants of the DB-outlier notion are proposed in [3] [8] [29] [40] [43] basically using the distances to the k-nearest neighbors (kNNs) or aggregates thereof.”
Statistical Analysis of Nearest Neighbor Methods for Anomaly Detection (2019, methodology): “Two versions of the NN anomaly detection algorithms have been proposed: kNN [20] and kNN [6].”
MAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks (2019, methodology): “he distance-based methods, a popular approach is the K-Nearest Neighbor (KNN) algorithm which computes the average distance to its k nearest neighbours and obtains anomaly scores based on this distane[17]. The ClusteringBased Local Outlier Factor (CBLOF) method is another example of the distance-bas”
Hierarchical Density Estimates for Data Clustering, Visualization, and Outlier Detection (2015, methodology): “Conceptually, for example, when referring to an estimate f of a given (possibly multivariate) PDF and a density threshold λ, Wishart observed that “...if f has two or more modes at the level of probability λ, then the covering will be partitioned into two or more disjoint connected subsets of points”
LoOP: local outlier probabilities (2009, methodology): “As a baseline, we used k-NN based outlier detection as defined in [9] (distance to k-th nearest neighbor, or“kNN” in short) and [2] (sum of the distances to the k nearest neighbors, or “kNN weight” in short).”
Outlier Analysis (2013, methodology): “In the context of high-dimensional data, there are two distinct lines of research, one of which investigates the efficiency of high dimensional outlier detection [46, 185, 467], and the other investigates the more fundamental issue of the effectiveness of high dimensional outlier detection [4, 273].”
Anomaly detection in dynamic networks: a survey (2015, methodology): “Vertex detection [55, 56, 67, 106] Subgraph detection [3, 13, 30, 31] Change detection [38] Edge detection [22] Change detection [112] Vertex detection [129] Event detection [68, 72, 75, 77, 83, 84, 93, 113, 114, 115, 122] Change detection [7, 65, 66] Edge detection [1, 80] Subgraph detection [60, 8”