Challenges 2019
During this second session, seven other challenges from the Challenge data website are presented :
- " Prediction of spatiotemporal PM10 fine particle concentration " presented by Grégoire Jauvion from Plume Labs. The aim of the challenge is to predict PM10 levels measured by certain air quality monitoring stations, based on measurements provided by adjacent monitoring stations and certain urban characteristics.
- " Predicting rat brain activity from temporal patterns of action potentials " presented by Ilya Prokin of the Group for Neural Theory (GNT) at ENS. The aim of the challenge is to predict the state of a rat's brain activity based on the time of occurrence of action potentials within a given hippocampal neuron.
- " Predicting electricity consumption for the pricing of electricity supplied " presented by Alexis Lucido from BCM Energy. The aim of the challenge is to predict the electricity consumption of two potential customers over one year, based on a set of consumption histories of other customers with similar characteristics but geographically distant, as well as geographical and meteorological data.
- " Screening and diagnosis of esophageal cancer using in vivo images presented by Fanny Louvet-de Verchère of Mauna Kea Technologies. The aim of the challenge is to classify endoscopic images of the esophagus into 4 classes representing different stages of esophageal cancer.
- " Predicting the loss experience of a building " presented by Clémence Devries from Generali. The aim of the challenge is to predict the water damage claims experience of an insured building over a one-year period, based on its characteristics.
- " Classification and optimization of quality of life at work presented by Sylvain Le Corff from Oze-Energies. The aim of the challenge is to predict the comfort felt by building occupants, based on environmental data measured in real time by sensors within the building.
- " Pricing exotic options by multidimensional nonlinear interpolation " presented by Olivier Croissant from Natixis. The aim of the challenge is to predict the value of exotic options contained in debt securities, based on 23 parameters characterizing them.
