Full text
| |
| Author(s): |
Lucas Oliveira Batista
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Instituto de Computação |
| Defense date: | 2015-07-07 |
| Examining board members: |
Claudia Maria Bauzer Medeiros;
Carla Geovana do Nascimento Macario;
Ariadne Maria Brito Rizzoni Carvalho
|
| Advisor: | Claudia Maria Bauzer Medeiros |
| Abstract | |
Time series are used in several knowledge domains, such as economics, meteorology and agriculture. In many situations, scientists often associate annotations to series during their analysis. They have moreover to search and correlate many types of time series in order to conduct their research. This process is hampered not only by the heterogeneity among the series, but also by the search for relevant series to compute a given correlation. The predominant methods to search series are based either in keyword (annotation) matching or in pattern matching. They do not support looking for additional information not always directly associated to series. Given this scenario, this dissertation proposes TS³Annotation, a framework based on the use of semantic annotations to support the study of correlations among time series. The main contributions of this work are: (1) a time series semantic annotation model; (2) and the TS³Annotation a framework that allows experts to create semantic annotations and uses these annotations as the basis for the search (AU) | |
| FAPESP's process: | 14/07303-1 - NavScales: navigating through scales in space, time and knowledge domain |
| Grantee: | Lucas Oliveira Batista |
| Support Opportunities: | Scholarships in Brazil - Master |