Model Evaluation, Model Selection, and Algorithm Selection in Machine Learning

Abstract

The correct use of model evaluation, model selection, and algorithm selection techniques is vital in academic machine learning research as well as in many industrial settings. This article reviews different techniques that can be used for each of these three subtasks and discusses the main advantages and disadvantages of each technique with references to theoretical and empirical studies. Further, recommendations are given to encourage best yet feasible practices in research and applications of machine learning. Common methods such as the holdout method for model evaluation and selection are covered, which are not recommended when working with small datasets. Different flavors of the bootstrap technique are introduced for estimating the uncertainty of performance estimates, as an alternative to confidence intervals via normal approximation if bootstrapping is computationally feasible. Common cross-validation techniques such as leave-one-out cross-validation and k-fold cross-validation are reviewed, the bias-variance trade-off for choosing k is discussed, and practical tips for the optimal choice of k are given based on empirical evidence. Different statistical tests for algorithm comparisons are presented, and strategies for dealing with multiple comparisons such as omnibus tests and multiple-comparison corrections are discussed. Finally, alternative methods for algorithm selection, such as the combined F-test 5x2 cross-validation and nested cross-validation, are recommended for comparing machine learning algorithms when datasets are small.

arXiv commentv3 (Nov 2020): Fixes SD from pooled proportions in Sec 4.2 Fixes exact binomial p-value in Sec 4.4 by using max(B, C) instead of B in the sum. Fixes typo using wrong symbols in Looney's F-test's SSA in Sec 4.7

References in corpus (1)

• Scikit-learn: Machine Learning in Python

Cited by in corpus (19)

• Machine learning and AI-based approaches for bioactive ligand discovery and GPCR-ligand recognition
• A Survey on Data-driven Software Vulnerability Assessment and Prioritization
• Review: Deep Learning in Electron Microscopy
• Weighted Random Search for CNN Hyperparameter Optimization
• Recognizing Developers' Emotions while Programming
• A Hierarchy of Limitations in Machine Learning
• A Question-answering Based Framework for Relation Extraction Validation
• MHAttnSurv: Multi-Head Attention for Survival Prediction Using Whole-Slide Pathology Images
• Learning-based Prediction and Uplink Retransmission for Wireless Virtual Reality (VR) Network
• Better Model Selection with a new Definition of Feature Importance
• Visual Framing of Science Conspiracy Videos: Integrating Machine Learning with Communication Theories to Study the Use of Color and Brightness
• Learn Like The Pro: Norms from Theory to Size Neural Computation
• 3D medical image segmentation with labeled and unlabeled data using autoencoders at the example of liver segmentation in CT images
• Learning the hypotheses space from data through a U-curve algorithm
• A scale invariant ranking function for learning-to-rank: a real-world use case
• Modeling User Empathy Elicited by a Robot Storyteller
• Towards Interpretable Reasoning over Paragraph Effects in Situation
• Speech Representations and Phoneme Classification for Preserving the Endangered Language of Ladin
• Uncertainty Estimation For Community Standards Violation In Online Social Networks