FDMI-AMG :: Mining of massive and uncertain data: Contribution of gradual patterns
In charge LIMOS : MEPHU NGUIFO EngelbertCoordinator : MEPHU NGUIFO Engelbert
Start of project : Sept. 1, 2022 - End of project : Aug. 31, 2024
Project driven by LIMOS
This project is part of the cooperation between the University of Clermont Auvergne (France) and the University of Yaoundé I (Cameroon) and its main goal is the extraction of gradual patterns in the presence of massive and uncertain data. Gradual patterns are an approach of extracting covariations as plus/minus x, plus/minus y. Several questions will be studied:
1- The correlation study task will be done through a comparative theoretical and experimental study between the correlations of the pairs of attributes captured by the gradual patterns and those detected using the statistical measures and then propose a measure of calculation of correlations between several attributes .
2- The task of pattern extraction in a multi-view or uncertain context is important in the context of massive data, and this double aspect can be taken into account during the preprocessing phase, or postprocessing or during the extraction of gradual patterns.
The expected results are the proposal of gradual pattern extraction methods for uncertain data, for multi-view data; studying the relevance of graded patterns to real-world application cases such as recommendation, and in mining COVID-19 data.
Partner Organizations :
UY1 |
Financier : CNRS