Ouliz2017

Reference

Ouliz, R. (2017) Prédiction de la variabilité spatiale de la disponibilité en biomasse résiduelle à l'aide de l'apprentissage automatique. Master's thesis, Université Laval. (URL )

Abstract

Residual forest biomass is the woody biomass left over on the forest floor after harvesting. This biomass can be used as a source of renewable energy, at a price that may be, under certain conditions, competitive relative to other energy sources. The success of the use of residual forest biomass depends in part on an effective management of its supply chain. Thus, the risk management of supply disruption of residual forest biomass is essential to ensure the potential for expansion of a customer’s distribution network. This project aims to improve the supply chain profitability of residual forest biomass through effective management of sources of error related to the estimation of the availability of biomass. This is the estimation of the spatial variability of residual biomass with acceptable accuracy by using machine learning techniques. Machine learning is an attempt to replicate the concept of learning. It consists to design algorithms capable to learn from examples or samples in order to predict the values of targets In our case study, the KNN method will allow us to estimate residual biomass of the target area units (polygons) from the k nearest neighbour plots. To this effect, we will estimate initially the spatial variability in the availability of residual biomass using the machine learning method KNN (k nearest neighbours). We then determine the error of our estimation using a bootstrap method. Finally, we will develop the location of the residual forest biomass quantity taking into account the estimation error. The estimation results obtained in the framework of this research indicate an accuracy of 59,5 % to 71 % centred around 65,4 % with an estimation error of 29 % to 34,5 %. Our methodology has yielded relevant results compared with the study of Bernier et al. (2010) which has had accuracy of estimation equal to 19% of forest biomass volume using the KNN method. The use of this method may also be relevant for estimating the commercial forest biomass and for the prediction of forest biomass of each tree species.

EndNote Format

You can import this reference in EndNote.

BibTeX-CSV Format

You can import this reference in BibTeX-CSV format.

BibTeX Format

You can copy the BibTeX entry of this reference below, orimport it directly in a software like JabRef .

@MASTERSTHESIS { Ouliz2017,
    TITLE = { Prédiction de la variabilité spatiale de la disponibilité en biomasse résiduelle à l'aide de l'apprentissage automatique },
    AUTHOR = { Ouliz, R. },
    SCHOOL = { Université Laval },
    YEAR = { 2017 },
    NOTE = { CEFTMS, Raulier, F. },
    ABSTRACT = { Residual forest biomass is the woody biomass left over on the forest floor after harvesting. This biomass can be used as a source of renewable energy, at a price that may be, under certain conditions, competitive relative to other energy sources. The success of the use of residual forest biomass depends in part on an effective management of its supply chain. Thus, the risk management of supply disruption of residual forest biomass is essential to ensure the potential for expansion of a customer’s distribution network. This project aims to improve the supply chain profitability of residual forest biomass through effective management of sources of error related to the estimation of the availability of biomass. This is the estimation of the spatial variability of residual biomass with acceptable accuracy by using machine learning techniques. Machine learning is an attempt to replicate the concept of learning. It consists to design algorithms capable to learn from examples or samples in order to predict the values of targets In our case study, the KNN method will allow us to estimate residual biomass of the target area units (polygons) from the k nearest neighbour plots. To this effect, we will estimate initially the spatial variability in the availability of residual biomass using the machine learning method KNN (k nearest neighbours). We then determine the error of our estimation using a bootstrap method. Finally, we will develop the location of the residual forest biomass quantity taking into account the estimation error. The estimation results obtained in the framework of this research indicate an accuracy of 59,5 % to 71 % centred around 65,4 % with an estimation error of 29 % to 34,5 %. Our methodology has yielded relevant results compared with the study of Bernier et al. (2010) which has had accuracy of estimation equal to 19% of forest biomass volume using the KNN method. The use of this method may also be relevant for estimating the commercial forest biomass and for the prediction of forest biomass of each tree species. },
    URL = { https://corpus.ulaval.ca/jspui/handle/20.500.11794/27789 },
    TIMESTAMP = { 2019-10-09 },
}

********************************************************** *************************** FRQNT ************************ **********************************************************

Un regroupement stratégique du

********************************************************** *********************** Infolettre *********************** **********************************************************

Abonnez-vous à
l'Infolettre du CEF!

********************************************************** ***************** Pub - Congrès Mycelium ****************** **********************************************************

Reporté en 2021

********************************************************** ***************** Pub - IWTT ****************** **********************************************************

Reporté en 2021

**********************************************************

***************** Pub - Symphonies_Boreales ****************** **********************************************************

********************************************************** ***************** Boîte à trucs *************** **********************************************************

CEF-Référence
La référence vedette !

Jérémie Alluard (2016) Les statistiques au moments de la rédaction 

  • Ce document a pour but de guider les étudiants à intégrer de manière appropriée une analyse statistique dans leur rapport de recherche.

Voir les autres...