Home / Current Issue / Paper 1719157
Detection of Semantic Events Using Changepoint and Ontology
Subject area: Science,Engineering and Technology · Area of research: Oceanography
DOI: https://doi.org/10.64388/IREV9I12-1719157
Abstract
With increased availability of time series data of oceanographic and environmental variables, there is a necessity for developing a way to recognize events out of continuous observation data. To solve the problem described above, this research offers a two-stages framework to recognize semantic events utilizing change point detection and ontology-based reasoning approach. At the first stage, change point detection techniques are used to identify significant events out of raw time series data. Pruned Exact Linear Time (PELT) algorithm is applied to find points of changes in statistical characteristics of data, like mean or variance. Detected change points divide time series into homogeneous parts, each of which can be considered as an event. At the second stage, semantic interpretation of detected parts is carried out based on ontology. Knowledge about domain is represented in the form of ontology that describes relationships between parameters of the system, such as threshold value, trend and environmental conditions. In accordance with described characteristics, each part is classified to meaningful groups of normal and risk conditions. For this project, I am using the Pelt approach.
How to cite this paper
@article{1719157,
author = {Dr. V. Seshasrinivas, V. Thanuja},
title = {Detection of Semantic Events Using Changepoint and Ontology},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {12},
pages = {2482-2492},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1719157.pdf},
abstract = {With increased availability of time series data of oceanographic and environmental variables, there is a necessity for developing a way to recognize events out of continuous observation data. To solve the problem described above, this research offers a two-stages framework to recognize semantic events utilizing change point detection and ontology-based reasoning approach. At the first stage, change point detection techniques are used to identify significant events out of raw time series data. Pruned Exact Linear Time (PELT) algorithm is applied to find points of changes in statistical characteristics of data, like mean or variance. Detected change points divide time series into homogeneous parts, each of which can be considered as an event. At the second stage, semantic interpretation of detected parts is carried out based on ontology. Knowledge about domain is represented in the form of ontology that describes relationships between parameters of the system, such as threshold value, trend and environmental conditions. In accordance with described characteristics, each part is classified to meaningful groups of normal and risk conditions. For this project, I am using the Pelt approach.},
month = {June},
doi = {https://doi.org/10.64388/IREV9I12-1719157}
}