Item


Conformal prediction of air pollution concentrations for the Barcelona Metropolitan Region

This thesis is aimed to introduce a newly developed machine learning method, conformal predictors, for air pollution assessment. For the given area of study, the Barcelona Metropolitan Region (BMR), several conformal prediction models have been developed. These models use the specification which is called ridge regression confidence machine (RRCM). The conformal predictors that have been developed for the purposes of the present study are ridge regression models, and they always provide valid predictions. Instead of a point prediction, a conformal predictor outputs a prediction set, which is usually an interval. It is desired that these sets would be as small as possible. The underlying algorithm for the conformal predictors derived in this thesis is ordinary kriging. A kriging-based conformal predictor can capture spatial distribution of the data with the use of so-called "kernel trick"

Aquest treball est脿 destinat a introduir el nou m猫tode de les m脿quines d鈥檃prenentatge, els predictors de conformaci贸, per l鈥檃valuaci贸 de la contaminaci贸 de l鈥檃ire a la Regi贸 Metropolitana de Barcelona (RMB). Es fa servir l鈥檈specificaci贸 anomenada m脿quina de confian莽a de la regressi贸 cresta (RRCM). Els predictors de conformaci贸 que s鈥檋an desenvolupat per les finalitats d鈥檃quest estudi s贸n uns models de regressi贸 cresta, que sempre ofereixen prediccions v脿lides. Un predictor de conformaci贸 genera un conjunt de predicci贸, que 茅s gaireb茅 sempre un interval, i la intenci贸 茅s que sigui el m茅s petit possible. L鈥檃lgorisme subjacent dels predictors de conformaci贸 derivats i discutits al llarg d鈥檃questa tesi 茅s el kriging. El predictor de conformaci贸 basat en el kriging ordinari pot capturar la distribuci贸 espacial mitjan莽ant una t猫cnica que es diu "el truc del nucli" ("kernel trick")

Universitat de Girona

Manager: S谩ez Zafra, Marc
Other contributions: Universitat de Girona. Departament d鈥橢conomia
Author: Ivina, Olga
Date: 2012 November 20
Abstract: This thesis is aimed to introduce a newly developed machine learning method, conformal predictors, for air pollution assessment. For the given area of study, the Barcelona Metropolitan Region (BMR), several conformal prediction models have been developed. These models use the specification which is called ridge regression confidence machine (RRCM). The conformal predictors that have been developed for the purposes of the present study are ridge regression models, and they always provide valid predictions. Instead of a point prediction, a conformal predictor outputs a prediction set, which is usually an interval. It is desired that these sets would be as small as possible. The underlying algorithm for the conformal predictors derived in this thesis is ordinary kriging. A kriging-based conformal predictor can capture spatial distribution of the data with the use of so-called "kernel trick"
Aquest treball est脿 destinat a introduir el nou m猫tode de les m脿quines d鈥檃prenentatge, els predictors de conformaci贸, per l鈥檃valuaci贸 de la contaminaci贸 de l鈥檃ire a la Regi贸 Metropolitana de Barcelona (RMB). Es fa servir l鈥檈specificaci贸 anomenada m脿quina de confian莽a de la regressi贸 cresta (RRCM). Els predictors de conformaci贸 que s鈥檋an desenvolupat per les finalitats d鈥檃quest estudi s贸n uns models de regressi贸 cresta, que sempre ofereixen prediccions v脿lides. Un predictor de conformaci贸 genera un conjunt de predicci贸, que 茅s gaireb茅 sempre un interval, i la intenci贸 茅s que sigui el m茅s petit possible. L鈥檃lgorisme subjacent dels predictors de conformaci贸 derivats i discutits al llarg d鈥檃questa tesi 茅s el kriging. El predictor de conformaci贸 basat en el kriging ordinari pot capturar la distribuci贸 espacial mitjan莽ant una t猫cnica que es diu "el truc del nucli" ("kernel trick")
Format: application/pdf
Document access: http://hdl.handle.net/10803/108341
Language: eng
Publisher: Universitat de Girona
Subject: Economia
Geologia. Meteorologia
Title: Conformal prediction of air pollution concentrations for the Barcelona Metropolitan Region
Type: doctoralThesis
Repository: TDX

Subjects

Authors