Mohamed Laib ~ $
- There are lies, damned lies, and statistics - -popularised by Mark Twain-
I am interested in data science and knowledge extraction from data. My current work is about developing new tools for mining space-time data using several statistical tools and machine learning algorithms. My research aims at exploring, understanding, and investigating challenging datasets. My work focuses on some important problems that can be faced when dealing with challenging data. Among the proposed methodologies:
Network analysis to investigate the wind monitoring system.
Spatial mapping of multifractal parameters obtained from each wind speed time series.
Spatial modelling of extreme wind speed distribution using extreme learning machine.
Unsupervised feature selection to reduce the existing redundancy in the input space.
Spatial mapping of the multifractal spectrum width. In this work, we can clearly distinguish three wind behaviours in Switzerland according to the topography of this country.
MFDFA results applied to a wind speed time series. This figure is obtained by using the MFDFA R library.
Detected communities using a mutual information-based network on wind speed monitoring system in Switzerland.More Publications
I have a great interest in R programming, I propose some libraries that I developed in collaboration with my colleagues. Please do not hesitate to contact me if you face problems with these libraries.
MFDFA: an R package computes the multifractal detrended fluctuation analysis for a time series. Available on GitHub.
SFtools: an R package for the unsupervised feature selection based on a space filing measure. Available on CRAN and GitHub.
Scene classification is an important and challenging problem in Earth observation remote sensing. This work, was partly presented at the EGU 2018 in the session Learning from spatial data: unveiling the geo-environment through quantitative approaches. It aims at exploring small-scale convolutional neural networks architecture to obtain good performance. For more information, please visit the website of this project, in which you find some Python codes and datasets used for this work:
Please do not hesitate to contact me for any question related to my work.Keep in touch