Ensemble Methods in Machine Learning

By Thomas G. Dietterich. 2007

Subfields: Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Probabilistic & Bayesian Methods

Citations in the corpus

Diversity With Cooperation: Ensemble Methods for Few-Shot Classification (2019, methodology): “To gain accuracy from averaging, various randomization or data augmentation techniques are typically used to encourage a high diversity of predictions [3, 6].”
Stacking with Auxiliary Features for Visual Question Answering (2018, background): “Ensembling multiple systems is a well known standard approach to improving accuracy in machine learning (Dietterich, 2000).”
Overcoming Multi-Model Forgetting (2019, neutral): “The use of multiple models to solve a single task dates back to model ensembles (Dietterich, 2000).”
Meta-Learning: A Survey (2018, neutral): “Note that while multi-task learning (Caruana, 1997) (learning multiple related tasks simultaneously) and ensemble learning (Dietterich, 2000) (building multiple models on the same task), can often be meaningfully combined with meta-learning systems, they do not in themselves involve learning from pr”
Sequence Student-Teacher Training of Deep Neural Networks (2016, methodology): “To achieve large combination gains, the systems in the ensemble need to be both diverse and individually accurate [5].”
Machine learning in manufacturing: advantages, challenges, and applications (2016, background): “The committee or ensemble contains a number of base learners like NNs, trees, or nearest neighbor (Dietterich, 2000; Opitz & Maclin, 1999).”
Blending LSTMs into CNNs (2015, background): “It is interesting to combinte these two types of models because they seem to meet the conditions of Dietterich (2000): “a necessary and sufﬁcient condition for an ensemble of classiﬁers to be more accurate than any of its individual members is if the classiﬁers are accurate and diverse".”
Seeing the Forest Through the Trees (2007, methodology): “The main goal in the design of ensemble methods is to increase the predictive accuracy of the classifier and studies indeed have shown the discrimination power of ensemble methods both theoretically and empirically [1, 4, 7], and in propositional as well as relational learning [12, 10, 6, 18].”
Mixture of experts: a literature survey (2012, unknown): “Hansen (1999, 2000) reported a potential problem that may occur in the competitive ME learning procedure, the zero co-efﬁcient problem.”
Automated Machine Learning: Methods, Systems, Challenges (2019, unknown): “Checkpoints Ensemble Adapting the ideas from [12, 18, 19] and capitalizing on our distributed framework, we simultaneously train several instances of our algorithm with different sets of hyperparameters, and then pick the best checkpoints in accordance to validation runs on 64 random seeds.”
A comprehensive comparison of random forests and support vector machines for microarray-based cancer classification (2008, background): “Specifically, if the generative linear decision function is not orthogonal to the coordinate axes, then a decision tree of infinite size is required to represent this function without error [17].”
Variable selection using random forests (2010, background): “fr (Christine Tuleau-Malot) ple Dietterich (1999) and Dietterich (2000)). Let us briefly recall the statistical framework by considering a learning set L = {(X1,Y1), . . . , (Xn,Yn)} made ofn i.i.d. observations of a random vector ( X,Y). VectorX = (X1, ...,Xp) contains predictors or explanatory var”
Endowing a Content-Based Medical Image Retrieval System with Perceptual Similarity Using Ensemble Strategy (2016, methodology): “The work of Bosis [29] relies on a semi-automated segmentation process, where the resulting images are represented through texture features and then classified by an ensemble of artificial neural networks (ANNs) [30], achieving varying results for evaluated Bfatty^ and Bdense^ classes.”
Models for early prediction of at-risk students in a course using standards-based grading (2016, background): “Usually ensemble models have lower error in predicting future data than any of the individual models that make them up (Dietterich, 2000).”
Machine learning and deep learning—A review for ecologists (2022, background): “…the complexity of an ensemble model is typically related to the average complexity of its ensemble members and the differences between them, which in turn affect error and variance of the ensemble estimator (Bernardo & Smith, 2009 ; Dietterich, 2000 ; Dormann et al., 2018; Ganaie et al., 2021 ).”