Model sparsity and brain pattern interpretation of classification models in neuroimaging

Rasmussen, Peter Mondrup; Hansen, Lars Kai; Madsen, Kristoffer Hougaard; Churchill, Nathan W.; Strother, Stephen C.

doi:10.1016/j.patcog.2011.09.011

Cited by 114 publications

(119 citation statements)

References 59 publications

Supporting

Mentioning

111

Contrasting

Order By: Relevance

“…However, sparsity alone is not sufficient for making reasonable and stable inferences, in cases of very high voxel space and small number of training samples. In such cases, the plain sparse learning models such as the ℓ 1 -norm regularized model also called LASSO [22], often provide overly sparse and hard to interpret solutions [23]. Thus we need to extend the plain sparse learning model to incorporate some important structural features of brain imaging data in order to achieve a more stable, reliable and interpretable support identification results.…”

Section: B Advantages and Limitations Of Sparsity Applied To Nueroimmentioning

confidence: 99%

Randomized Structural Sparsity-Based Support Identification with Applications to Locating Activated or Discriminative Brain Areas: A Multicenter Reproducibility Study

Wang

Zhang

Zheng

et al. 2015

IEEE Trans. Auton. Mental Dev.

View full text Add to dashboard Cite

show abstract

Section: B Advantages and Limitations Of Sparsity Applied To Nueroimmentioning

confidence: 99%

Randomized Structural Sparsity-Based Support Identification with Applications to Locating Activated or Discriminative Brain Areas: A Multicenter Reproducibility Study

Wang

Zhang

Zheng

et al. 2015

IEEE Trans. Auton. Mental Dev.

View full text Add to dashboard Cite

show abstract

“…A recent work [8] studies the impact of model selection on the reproducibility and stability of the estimated models for simpler learning methods. In the present work, we select the hyper-parameters in the usual way of maximizing the classification accuracy over the internal LOSO-CV and we assess the stability of the resulting models in the external LOSO-CV.…”

Section: Experimental Protocol and Assessmentmentioning

confidence: 99%

“…However, often these models provide overly sparse solutions, or activation patterns, where the non-zero coefficients are assigned to disparate regions across the brain, without exploiting any spatial or temporal prior information [3], [4], [8].…”

Section: Introductionmentioning

confidence: 99%

Structured Sparsity Models for Brain Decoding from fMRI Data

Baldassarre

Mourão-Miranda

Pontil

2012

2012 Second International Workshop on Pattern Recognition in NeuroImaging

View full text Add to dashboard Cite

Abstract-Structured sparsity methods have been recently proposed that allow to incorporate additional spatial and temporal information for estimating models for decoding mental states from fMRI data. These methods carry the promise of being more interpretable than simpler Lasso or Elastic Net methods. However, despite sparsity has often been advocated as leading to more interpretable models, we show that by itself sparsity and also structured sparsity could lead to unstable models.We present an extension of the Total Variation method and assess several other structured sparsity models on accuracy, sparsity and stability. Our results indicate that structured sparsity via the Sparse Total Variation can mitigate some of the instability inherent in simpler sparse methods, but more research is required to build methods that can reliably infer relevant activation patterns from fMRI data.

show abstract

“…Almost all collaborative filtering algorithms come with one or more adjustable hyperparameters, such as regularization and learning rate. The setting of these values plays a key role on the generalizability and accuracy of the predictive model [20]. The problem of finding better hyperparameter values for an algorithm to improve prediction accuracy or to optimize certain goals is called hyperparameter optimization or model selection.…”

Section: Related Workmentioning

confidence: 99%

Continuous hyperparameter optimization for large-scale recommender systems

Chan

Treleaven

Capra

2013

2013 IEEE International Conference on Big Data

View full text Add to dashboard Cite

Abstract-While the prediction accuracy of a large-scale recommender system can generally be improved by learning from more and more training data over time, it is unclear how well a fixed predictive model can handle the changing business dynamics in a real-world scenario. The adjustment of a predictive model is controlled by the hyperparameter settings of a selected algorithm. Although the problem of hyperparameter optimization has been studied for decades in various disciplines, the adaptiveness of the initially selected model is not as well understood. This paper presents an approach to continuously re-select hyperparameter settings of the algorithm in a large-scale retail recommender system. In particular, an automatic hyperparameter optimization technique is applied on collaborative filtering algorithms in order to improve prediction accuracy. Experiments have been conducted on a large-scale real retail dataset to challenge traditional assumption that a one-off initial hyperparameter optimization is sufficient. The proposed approach has been compared with a baseline approach and a widely used approach with two scalable collaborative filtering algorithms. The evaluations of our experiments are based on a 2-year real purchase transaction dataset of a large retail chain business, both its online e-commerce site and its offline retail stores. It is demonstrated that continuous hyperparameter optimization can effectively improve the prediction accuracy of a recommender system. This paper presents a new direction in improving the prediction performance of a large-scale recommender system.

show abstract

Model sparsity and brain pattern interpretation of classification models in neuroimaging

Cited by 114 publications

References 59 publications

Randomized Structural Sparsity-Based Support Identification with Applications to Locating Activated or Discriminative Brain Areas: A Multicenter Reproducibility Study

Randomized Structural Sparsity-Based Support Identification with Applications to Locating Activated or Discriminative Brain Areas: A Multicenter Reproducibility Study

Structured Sparsity Models for Brain Decoding from fMRI Data

Continuous hyperparameter optimization for large-scale recommender systems

Contact Info

Product

Resources

About