Skip to Main content Skip to Navigation
New interface
Conference papers

Multi-factor Prediction and Parameters Identification based on Choquet Integral: Smart Farming Application

Abstract : In this paper, we consider the domain of smart farming aiming at agronomic processes optimization, and, more particularly, the issue of predicting the growth stages transitions of a plant. As existing automated predictions are not accurate nor reliable enough to be used in the farming process, we propose here an approach based on Choquet integral, enabling the passage from multiple imperfect predictions to a more accurate and reliable one, considering the relevance of each source in the prediction as well as the interactions, synergies, or redundancies between factors. Identifying the parameter values defining a Choquet-based decision model being not straightforward, we propose an approach based on an observation history. Our proposal defines an evaluation function assigning to any potential solution a predictive capability, quantifying a degree of order present in its output, and an associated optimisation process based on truth degrees regarding a set of inequalities. A case study concerns smart farming, the prototype we implemented enabling, for a given culture and several input sources, to help farmers to predict the next growth stage. The experimental results are very encouraging, the predicted day remaining stable despite presence of noise on evidence values.
Document type :
Conference papers
Complete list of metadata

https://hal-cnam.archives-ouvertes.fr/hal-03789592
Contributor : Marie-Liesse Bertram Connect in order to contact the contributor
Submitted on : Tuesday, September 27, 2022 - 3:20:48 PM
Last modification on : Thursday, October 27, 2022 - 12:13:36 PM

Links full text

Identifiers

Citation

Yann Pollet, Jérôme Dantan, Hajer Baazaoui. Multi-factor Prediction and Parameters Identification based on Choquet Integral: Smart Farming Application. 17th International Conference on Software Technologies, Jul 2022, Lisbon, Portugal. pp.340-348, ⟨10.5220/0011317900003266⟩. ⟨hal-03789592⟩

Share

Metrics

Record views

57