Detection of Sequential Outliers Using a Variable Length Markov Model

The problem of mining for outliers in sequential datasets is crucial to forward appropriate analysis of data. Therefore, many approaches for the discovery of such anomalies have been proposed. However, most of them use a sample of known typical sequences to build the model. Besides, they remain gree...

Full description

Saved in:
Bibliographic Details
Published in2008 Seventh International Conference on Machine Learning and Applications pp. 571 - 576
Main Authors Low-Kam, C., Laurent, A., Teisseire, M.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2008
Subjects
Concentration Inequality
Data analysis
DNA
Genetic mutations
Information Criterion
Machine learning
Outliers
Proteins
Robots
Sequences
Sequential Databases
Size measurement
Testing
Online AccessGet full text

Cover

More Information
Summary:The problem of mining for outliers in sequential datasets is crucial to forward appropriate analysis of data. Therefore, many approaches for the discovery of such anomalies have been proposed. However, most of them use a sample of known typical sequences to build the model. Besides, they remain greedy in terms of memory usage. In this paper we propose an extension of one such approach, based on a probabilistic suffix tree and on a measure of similarity. We add a pruning criterion which reduces the size of the tree while improving the model, and a sharp inequality for the concentration of the measure of similarity, to better sort the outliers. We prove the feasibility of our approach through a set of experiments over a protein database.
ISBN:0769534953
9780769534954
DOI:10.1109/ICMLA.2008.137