Title: Feature Construction for Multiclass Classification in Remote Sensing
Speaker: João E. Batista, LASIGE – DI/FCUL
When: Wednesday, April 7th, 17h45
Short summary: Deforestation has serious implications on biodiversity, on rural communities that are dependent on the forest for survival and on greenhouse gas emissions that are driving the global climate. One way to control deforestation is through the use of mechanisms that promote sustainable forestry and support economic development, like REDD+ of the United Nations. However, these mechanisms require careful forest monitorization to be effective. Having this into consideration, this work attempts to improve the quality of the machine learning methods used for forest monitorization by promoting Genetic Programming as an automatic method for the creation of robust indices for multiclass classification in a pixel-level approach and, in current work, in a time-series approach.
Short Bio: João E. Batista is a 3rd year PhD student that has been working in the area of Genetic Programming (GP) since 2016, and in the area of Remote Sensing (RS) since 2019. In his previous work, GP was used in the context of symbolic regression, binary classification and multiclass classification, with implicit Feature Construction in each problem. His work in the area of RS dealt with the detection of burnt areas and with the classification of landcover types within satellite images, in a pixel-level approach.