MétaCan
Menu
Back to cohort
Record W7048938799

Méthodes d’estimation des quantiles conditionnels en
\nhydro-climatologie.

2017· dissertation· fr· W7048938799 on OpenAlexaboutno aff

Bibliographic record

VenueEspaceINRS Institutional Digital Repository (Institut National de la Recherche Scientifique) · 2017
Typedissertation
Languagefr
FieldPhysics and Astronomy
TopicLightning and Electromagnetic Phenomena
Canadian institutionsnot available
Fundersnot available
KeywordsQuantileDistribution (mathematics)Influence function
DOInot available

Abstract

fetched live from OpenAlex

Cette thèse de doctorat a pour objectif de développer de nouvelles méthodes pour l’estimation \ndes quantiles des événements hydrologiques extrêmes en présence des covariables climatiques (donc, \nil s’agit de l’estimation des quantiles conditionnels) dans le cas où la dépendance entre la variable \nd’intérêt et les covariables est de type non nécessairement linéaire ou inconnu. Dans cette thèse, \nnous proposons trois approches pour l’estimation des quantiles conditionnels : deux sont basées sur \ndes dépendances de type B-Splines décrivant soit la relation entre les paramètres d’une fonction \nde répartition d’une variable d’intérêt et les covariables, soit le lien entre la variable d’intérêt et \nles covariables sous un modèle de régression des quantiles et une approche basée sur des dépendances \nde type copule décrivant le lien entre une fonction de répartition d’une variable d’intérêt \net les covariables. Ces trois approches ont été tout d’abord comparées avec les approches classiques \nqui reposent sur des dépendances linéaires ou quadratiques et ensuite comparées entre elles \nafin de déterminer le meilleur estimateur. Aussi, elles ont été appliquées sur des bases de données \nhydro-climatiques afin d’estimer le risque des extrêmes dans certaines régions du monde, plus spécifiquement \nle nord de l’Afrique et le nord-est du Canada. Les résultats de nos travaux ont montré \nle grand avantage de nos approches par rapport aux méthodes classiques. En plus, l’estimation des \nquantiles conditionnels basée sur les copules a montré une grande performance par rapport aux deux autres approches basées sur les fonctions B-Splines. Cette performance a été démontrée en se \nbasant sur des simulations considérant différents modèles statistiques. En effet, ajouter un modèle \nde dépendance, par exemple une copule, aux modèles des quantiles conditionnels permet de capturer \nla structure de dépendance globale entre la variable d’intérêt et les covariables. Par conséquent, cela \npermet de diminuer largement le biais d’estimation, ce qui donne des estimations de risque de plus \nen plus précises pour la gestion des événements hydrologiques ou climatiques extrêmes. This PhD thesis aims to develop new methods for estimating the quantiles of extreme hydrological \nevents in the presence of climatic covariates (i.e., the estimation of conditional quantiles) in \nthe context where the dependence between the variable of interest and covariates is not necessarily \nlinear or known. In this thesis, we propose three approaches for estimating the conditional quantiles: \ntwo approaches are based on the B-Splines functions which describe either the relationship between \nthe parameters of a cumulative distribution function of a variable of interest and the covariables, or \nthe link between the variable of interest and the covariates under a quantile regression model and \nan approach based on copula functions which describe the link between a cumulative distribution \nfunction of a variable of interest and the covariables. We first compared these three approaches with \nmore classical non stationary approaches which are based on linear or quadratic dependency models \nand then compared the three proposed approaches to each other in order to determine the best estimators. \nAlso, they have been applied to case studies where hydro-climatic data are used to estimate \nthe risk of extremes in some regions of the world, specifically northern Africa and eastern Canada. \nThe results of our work have shown the great advantage of using our approaches compared to the \nclassical approaches and they showed the performance of estimating the conditional quantiles based \non the copula approach. This performance is demonstrated by simulations which consider different \nstatistical models. Indeed, using copula function to estimate conditional quantile models allows to \ncapture the overall dependence structure between the variable of interest and covariates and then provides an important estimation improvement.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.005
metaresearch head score (Gemma)0.016
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.988
Threshold uncertainty score0.064

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0050.016
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.002
Bibliometrics0.0010.001
Science and technology studies0.0000.001
Scholarly communication0.0020.001
Open science0.0020.002
Research integrity0.0010.003
Insufficient payload (model declined to judge)0.0190.006

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.077
GPT teacher head0.352
Teacher spread0.275 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designSimulation or modeling
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2017
Admission routes1
Has abstractyes

Explore more

Same venueEspaceINRS Institutional Digital Repository (Institut National de la Recherche Scientifique)Same topicLightning and Electromagnetic PhenomenaFrench-language works237,207