A Survey of Outlier Detection Methodologies

  title={A Survey of Outlier Detection Methodologies},
  author={Victoria J. Hodge and Jim Austin},
  journal={Artificial Intelligence Review},
Outlier detection has been used for centuries to detect and, where appropriate, remove anomalous observations from data. Outliers arise due to mechanical faults, changes in system behaviour, fraudulent behaviour, human error, instrument error or simply through natural deviations in populations. Their detection can identify system faults and fraud before they escalate with potentially catastrophic consequences. It can identify errors and remove their contaminating effect on the data set and as… 

Machine learning to detect anomalies in datacenter

This thesis investigates the possibility of using anomaly detection on performance data of virtual servers in a datacenter to detect malfunctioning servers. Using anomaly detection can potentially ...

Cyber Data Anomaly Detection Using Autoencoder Neural Networks

The detection methodology proposed uses an autoencoder neural network optimized via design of experiments for the identification of anomalous network events that can operate in conjunction with alternate tools for outlier detection, enhancing the overall anomaly detection capability of cyber analysts.

Ensemble based unsupervised anomaly detection

Using a combination of statistical and clustering approaches, an ensemble of algorithms provide automatic anomaly detection in an Application-to-person networking environment which can be scaled to different domains using hierarchical time series data.

Soybean mintill 2455 12 Soybean clean 593 13 Wheat 205 14 Woods 1265 15 Building Grass Tree Drives 386 16 Stone steel towers 93

This paper chose Indian Pines as the test site to classify it into 16 classes of tree species and compare them with ground truth to obtain the accuracy, using various algorithms with machine learning process, including LDA and SVM.

A Systematic Literature Review on Outlier Detection in Wireless Sensor Networks

The current paper presents an improved taxonomy of outlier detection techniques that will help researchers and practitioners to find the most relevant and recent studies related to outlier Detection in WSNs.

The link between self-esteem and social relationships: A meta-analysis of longitudinal studies.

The findings suggest that the link between people's social relationships and their level of self-esteem is truly reciprocal in all developmental stages across the life span, reflecting a positive feedback loop between the constructs.

Abnormal Signal Recognition with Time-Frequency Spectrogram: A Deep Learning Approach

A new abnormal signal recognition scheme, which com-bines time-frequency analysis with deep learning to effectively identify synthetic abnormal communication signals, which can effectively recognize the abnormal signals under various parameter configurations.

An Efficient Video Streaming Architecture with QoS Control for Virtual Desktop Infrastructure in Cloud Computing

An architecture including flexible QoS control to improve the users’ quality of experience (QoE) and a novel compression algorithm of 2D images, designed to guarantee the best image quality and to reduce video delay is presented.

LPRNet: A Novel Approach for Novelty Detection in Networking Packets

This paper has proposed a novel approach LPRNet, a Denoising Autoencoder which uses algorithms such as Least Trimmed Square, Projected Gradient Descent and Robust Principal Component Analysis, to solve the above-mentioned problems.

Methodology Maps for Model-Based Sensor-Data Interpretation to Support Civil-Infrastructure Management

Methodology maps are introduced in order to provide guidance for appropriate use of model-based methodologies to interpret monitoring information through reference to uncertainty conditions and computational constraints, which improves asset-management decision making.



Statistical Independence and Novelty Detection with Information Preserving Nonlinear Maps

This work applies the criterion of minimal mutual information to the real world problem of electrical motor fault detection treated as a novelty detection task, and generalizes to nonlinear transformations by only demanding perfect transmission of information.

Outlier detection and motion segmentation

We present a new method for solving the problem of motion segmentation, identifying the objects within an image moving independently of the background. We utilize the fact that two views of a static

Robust Regression and Outlier Detection

This paper presents the results of a two-year study of the statistical treatment of outliers in the context of one-Dimensional Location and its applications to discrete-time reinforcement learning.

UCI Repository of machine learning databases

A massively parallel architecture for a self-organizing neural pattern recognition machine

A System for the Analysis of Jet System Vibration Data

  • Integrated ComputerAided Engineering
  • 1999

A study of distance-based machine learning algorithms

It is shown that the k-nearest neighbor algorithm (kNN) outperforms the first nearest neighbor algorithm only under certain conditions, and methods for choosing the value of k for kNN are investigated, and two methods for learning feature weights for a weighted Euclidean distance metric are proposed.

Novelty detection and neural network validation

This paper provides a quantitative procedure for measuring novelty, and its performance is demonstrated using an application involving the monitoring of oil flow in multi-phase pipelines.

Outliers in Statistical Data