ODD 2014 - 2014 Workshop on Outlier Detection & Description under Data Diversity

We consider outlier mining as a three-fold challenge: outlier detection, outlier description, and outlier modeling under data diversity. Outlier detection refers to the quantitative mining of the outliers and tries to identify outliers from large datasets. On the other hand, outlier description concerns with the qualitative mining of outliers and tries to interpret or explain outlier properties. In addition to detection and description, one faces the curse of data diversity, with the increase in complexity of real-world datasets including temporal, spatial, sequence, graph, and multi-dimensional space anomalies. The topics of detection and description for diverse data are rarely considered in unison, and literature for these tasks is spread over different research communities. The main goal of ODD2 is to bridge this gap and provide a venue for knowledge exchange between these different research areas.
For today's outlier mining applications to be successful, three directions should be addressed: (i) how to create outlier detection models, (ii) how to serve the needs of certain applications with respect to interpretability, and (iii) how to create outlier models that can handle datasets consisting of diverse sources. We remark that those directions are not disjoint, in contrast they are quite intertwined: as the diversity and thus complexity of the data increases, outlier description not only becomes more challenging but also even more necessary.
Most of today's data sources are very heterogeneous: multi-/high-dimensional data points, evolving data streams, text, heterogeneous graphs, spatio-temporal data, and so on. For example consider social media platforms such as Twitter or Facebook. The diversity of data in such platforms is immense; ranging from relations among users (graphs), user demographics (high-dimensional features), user-generated content (text), temporal dynamics (data streams), heterogeneous relational data in the form of likes, shares, tags, and so on. Outliers in those applications often correspond to various types of anomalies ranging from fake celebrities, fake user accounts, (social) malware, page-like-as-a-service, and so on. As a result, it becomes critical to develop outlier models that can work with diverse datasets, provide high quality outlier detection, but also describe the reasons for outlier properties to the human user.
Topics of Interest
The workshop covers several aspects of outlier detection and description and of related research fields. A non-exhaustive list of topics of interest is given below:
Interleaved detection and description of outliers
Description models for given outliers
Pattern and local information based outlier description
Subspace outliers, feature selection, and space transformations
Ensemble methods for anomaly detection and description
Descriptive local outlier ranking
Identification of outlier rules
Finding intensional knowledge
Contextual and community outliers
Human-in-the-loop modeling and learning
Visualization techniques for interactive exploration of outliers
Comparative studies on outlier description
Related research fields
Contrast mining
Change and novelty detection
Causality analysis
Frequent itemset mining
Compression theory
Subgroup mining
Subspace learning
Formal outlier mining models
Supervised, semi-supervised, and unsupervised models
Statistical models
Distance-based models
Density-based models
Spectral models
Constraint-based models
Ensemble models
Outlier mining for complex databases
Graph data (e.g. community outliers)
Spatio-temporal data
Time series and sequential data
Online processing of stream data
Scalability to high dimensional data
Applications of outlier detection and description
Fraud in financial data
Intrusions in communication networks
Sensor network analysis
Social network analysis
Health surveillance
Customer profiling
... and many more ...
We encourage submissions describing innovative work in related fields that address the issue of diversity in outlier mining.