Análisis de series temporales mediante expresiones regulares temporizadas
Loading...
Official URL
Full text at PDC
Publication date
2024
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citation
Abstract
El objetivo principal de este TFG es mejorar ParetoLib, una herramienta existente de análisis de series temporales mediante la incorporación de Expresiones Regulares Temporales (Timed Regular Expression, TRE) para posteriormente aplicarlo a la detección de patrones de consumo eléctrico. Como resultado se mejorará esta búsqueda de patrones para detectar posibles consumos anómalos y poder detectar así diferentes ataques a infraestructuras eléctricas críticas como los Smart Grid. Para lograr este objetivo, se han integrado en la herramienta dos motores de análisis de TRE. Esto ha permitido estudiar diferentes tipos de expresiones regulares sobre una serie de datos previamente manipulados para obtener una expresión correcta que satisfaga un patrón de consumo eléctrico anómalo. Gracias a esto se ha podido asociar diferentes tipos de ataque con un comportamiento identificativo que ayude a estas infraestructuras a prevenir o detectar posibles consumos anómalos. También se ha mejorado la herramienta para que pueda interpretar de manera más visual los resultados obtenidos tras la ejecución de los motores de análisis TRE. Esta mejora tiene como objetivo facilitar la búsqueda de patrones de consumo, haciéndola aplicable a cualquier tipo de comportamiento. Para ello se ha llevado a cabo una fase de investigación previa, acompañada de una fase de implementación técnica y seguida de una fase de experimentación.
The main objective of this TFG is to improve ParetoLib, an existing time series analysis tool by incorporating Timed Regular Expressions (TRE) to later apply it to the detection of electricity consumption patterns. As a result, this pattern search will be improved to detect possible anomalous consumption and thus be able to detect different attacks on critical electrical infrastructures such as Smart Grids. To achieve this goal, two TRE analysis engines have been integrated into the tool. This has made it possible to study different types of regular expressions on a series of previously manipulated data to obtain a correct expression that satisfies an anomalous electricity consumption pattern. Thanks to this, it has been possible to associate different types of attack with an identifying behavior that helps these infrastructures to prevent or detect possible anomalous consumption. The tool has also been improved to allow a more visual interpretation of the results obtained after running the TRE analysis engines. This improvement is intended to facilitate the search for consumption patterns, making it applicable to any type of behavior. To this end, a preliminary research phase has been carried out, accompanied by a technical implementation phase and followed by an experimentation phase.
The main objective of this TFG is to improve ParetoLib, an existing time series analysis tool by incorporating Timed Regular Expressions (TRE) to later apply it to the detection of electricity consumption patterns. As a result, this pattern search will be improved to detect possible anomalous consumption and thus be able to detect different attacks on critical electrical infrastructures such as Smart Grids. To achieve this goal, two TRE analysis engines have been integrated into the tool. This has made it possible to study different types of regular expressions on a series of previously manipulated data to obtain a correct expression that satisfies an anomalous electricity consumption pattern. Thanks to this, it has been possible to associate different types of attack with an identifying behavior that helps these infrastructures to prevent or detect possible anomalous consumption. The tool has also been improved to allow a more visual interpretation of the results obtained after running the TRE analysis engines. This improvement is intended to facilitate the search for consumption patterns, making it applicable to any type of behavior. To this end, a preliminary research phase has been carried out, accompanied by a technical implementation phase and followed by an experimentation phase.
Description
Trabajo de Fin de Grado en Ingeniería del Software, Facultad de Informática UCM, Departamento de Sistemas Informáticos y Computación, Curso 2023/2024.