Regularisation (Regularization)
Last revised by Michael Woodburn on 12 Dec 2019
Citation, DOI, disclosures and article data
Citation:
Wang D, Woodburn M, Murphy A, et al. Regularisation (Regularization). Reference article, Radiopaedia.org (Accessed on 29 Mar 2024) https://doi.org/10.53347/rID-61718
Permalink:
rID:
61718
Article created:
15 Jul 2018,
David John Wang
Disclosures:
At the time the article was created David John Wang had no recorded disclosures.
View David John Wang's current disclosures
Last revised:
12 Dec 2019,
Michael Woodburn
Disclosures:
At the time the article was last revised Michael Woodburn had no recorded disclosures.
View Michael Woodburn's current disclosures
Revisions:
6 times, by
4 contributors -
see full revision history and disclosures
Sections:
Tags:
Regularisation is a process of reducing the complexity of a model through the inclusion of an additional parameter as in order to reduce the overfitting of a model to the training data.
In the context of radiology, a common model type used to interpret images is the convolutional neural network. Convolutional neural networks are intrinsically regularised to an extent due to the convolution of features at each layer. In practice, this trait forces the model to weight more features with intermediate magnitude rather than few features with high magnitude.
References
- 1. Stephen Marsland. Machine Learning. (2018) ISBN: 9781420067187
- 2. Goodfellow Ian, Yoshua Bengio and Aaron Courville. Deep learning. Cambridge, MA: The MIT Press, 2017.
Incoming Links
Articles:
Related articles: Artificial intelligence
- artificial intelligence (AI)
- imaging data sets
- computer-aided diagnosis (CAD)
- natural language processing
- machine learning (overview)
- visualising and understanding neural networks
- common data preparation/preprocessing steps
- DICOM to bitmap conversion
- dimensionality reduction
- scaling
- centring
- normalisation
- principal component analysis
- training, testing and validation datasets
- augmentation
- loss function
-
optimisation algorithms
- ADAM
- momentum (Nesterov)
- stochastic gradient descent
- mini-batch gradient descent
-
regularisation
- linear and quadratic
- batch normalisation
- ensembling
- rule-based expert systems
- glossary
- activation function
- anomaly detection
- automation bias
- backpropagation
- batch size
- computer vision
- concept drift
- cost function
- confusion matrix
- convolution
- cross validation
- curse of dimensionality
- dice similarity coefficient
- dimensionality reduction
- epoch
- explainable artificial intelligence/XAI
- feature extraction
- federated learning
- gradient descent
- ground truth
- hyperparameters
- image dataset normalisation
- image registration
- imputation
- iteration
- jaccard index
- linear algebra
- noise reduction
- normalisation
- R (Programming language)
- radiomics quality score (RQS)
- Python (Programming language)
- segmentation
- semi-supervised learning
- synthetic and augmented data
- overfitting
- underfitting
- transfer learning