Towards principled learner selection

Long learner evaluation times are no longer exceptional and often there is insufficient time to exhaustively test all candidate options. When deciding which learners to use, practitioners must rely on ad hoc testing and luck to identify the most accurate one. Given the importance of classification i...

Full description

Bibliographic Details
Main Author: Rattray, Andrew
Format: Thesis (University of Nottingham only)
Language:English
Published: 2013
Subjects:
Online Access:https://eprints.nottingham.ac.uk/13299/
_version_ 1848791700301938688
author Rattray, Andrew
author_facet Rattray, Andrew
author_sort Rattray, Andrew
building Nottingham Research Data Repository
collection Online Access
description Long learner evaluation times are no longer exceptional and often there is insufficient time to exhaustively test all candidate options. When deciding which learners to use, practitioners must rely on ad hoc testing and luck to identify the most accurate one. Given the importance of classification in decision making, this is unsatisfactory. Progress towards a principled approach requires accurate predictions of learner accuracy and evaluation time and this study examines the potential of traditional meta-learning approaches, with their emphasis on indirect explanatory variables, to deliver the required solutions. Here, 57 different indirect dataset characteristics, including those related to geometrical complexity, are used as explanatory variables, alongside sample-estimates, in building regression models of accuracy and time. The evidence presented firmly suggests that these indirect variables lack both the required predictive power and the time efficiency required for the development of practically useful models, and points instead towards basing the prediction of learner accuracy solely on sample-based models. The attempt at modelling learner evaluation time reveals some of the difficulties that this tough challenge presents.
first_indexed 2025-11-14T18:32:40Z
format Thesis (University of Nottingham only)
id nottingham-13299
institution University of Nottingham Malaysia Campus
institution_category Local University
language English
last_indexed 2025-11-14T18:32:40Z
publishDate 2013
recordtype eprints
repository_type Digital Repository
spelling nottingham-132992025-02-28T11:24:21Z https://eprints.nottingham.ac.uk/13299/ Towards principled learner selection Rattray, Andrew Long learner evaluation times are no longer exceptional and often there is insufficient time to exhaustively test all candidate options. When deciding which learners to use, practitioners must rely on ad hoc testing and luck to identify the most accurate one. Given the importance of classification in decision making, this is unsatisfactory. Progress towards a principled approach requires accurate predictions of learner accuracy and evaluation time and this study examines the potential of traditional meta-learning approaches, with their emphasis on indirect explanatory variables, to deliver the required solutions. Here, 57 different indirect dataset characteristics, including those related to geometrical complexity, are used as explanatory variables, alongside sample-estimates, in building regression models of accuracy and time. The evidence presented firmly suggests that these indirect variables lack both the required predictive power and the time efficiency required for the development of practically useful models, and points instead towards basing the prediction of learner accuracy solely on sample-based models. The attempt at modelling learner evaluation time reveals some of the difficulties that this tough challenge presents. 2013-07-15 Thesis (University of Nottingham only) NonPeerReviewed application/pdf en arr https://eprints.nottingham.ac.uk/13299/1/MRES_ARATTRAY_4145685_FINAL_CORRECTED.pdf Rattray, Andrew (2013) Towards principled learner selection. MRes thesis, University of Nottingham. machine learning algorithms
spellingShingle machine learning
algorithms
Rattray, Andrew
Towards principled learner selection
title Towards principled learner selection
title_full Towards principled learner selection
title_fullStr Towards principled learner selection
title_full_unstemmed Towards principled learner selection
title_short Towards principled learner selection
title_sort towards principled learner selection
topic machine learning
algorithms
url https://eprints.nottingham.ac.uk/13299/