Article

Article title SPECIAL TEMPORAL PATTERN RECOGNITION TECHNIQUE BASED ON HYBRID STOCHASTIC MODEL
Authors S.M. Kovalev, A.V. Sukhanov
Section SECTION II. ALGORITHMS AND SOFTWARE
Month, Year 04, 2014 @en
Index UDC 656.2 + 06
DOI
Abstract There is a wide introduction of automated information management systems based on databases and knowledge base in our days. Therefore there is a computer need in analysis of large information volumes received as a result of technical facilities and equipment work observing. Here various temporal data processing techniques are used for the identification and compilation of useful information. In particular there is time series clustering and classification techniques. This paper presents one of important problem decision in Data Mining dedicated to specific temporal patterns detection. Proposed technique based on unsupervised training of Markov chain model with productional “non-Markov” rules. Such approach could be used for wide problem decision because it is robust forthe lack of information. Represented experiments in one of time series standard sample implementation demonstrates relevance of such techniques for special pattern detection in temporal sets.

Download PDF

Keywords Anomaly detection; unsupervised learning; time series; Markov chain; productional models; special pattern recognition.
References 1. Grabocka J., Nanopoulos A., Schmidt-Thieme L. Invariant Time-Series Classification // European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD). – 2012. – P. 725-740.
2. Ковалев С.М., Гуда А.Н., Бутакова М.А. Гибридная стохастическая модель обнаружения особых типов паттернов в темпоральных данных // Вестник РГУПС. – 2013. – № 3 (51). – С. 36-42.
3. Суханов А.В. Стохастическая Марковская модель поиска аномалий в темпоральных данных // Труды Конгресса по интеллектуальным системам и информационным технологиям «IS&IT’13»: В 4 т. – М.: Физматлит, 2013. – Т. 1. – С. 177-181.
4. Chandola V., Banerjee A., Kumar V. Anomaly Detection: A Survey // ACM Computing Surveys, 2009. – Vol. 41(3). Article 15. – P. 1-72.
5. Ma J., Perkins S. Time-series novelty detection using one-class support vector machines // Proceedings of the International Joint Conference on Neural Networks. – July 2003. – Vol. 3. – P. 1741–1745.
6. Заде Л. Понятие лингвистической переменной и его применение к принятию приближенных решений. – М.: Мир, 1976. – 166 c.
7. Scheuner U. Fuzzy-Mengen Verknьpfung und Fuzzy-Arithmetik zur Sensordaten-Fusion // VDI-Verlag. – 2001. – Bd. 8.
8. Keogh E., Xi X., Wei L., Ratanamahatana C.A. The UCR time series classification/clustering homepage // http://www.cs.ucr.edu/ ~eamonn/time_series_data/.

Comments are closed.