Training radiomics-based CNNs for clinical outcome prediction: Challenges, strategies and findings

2022 ◽  
Vol 123 ◽  
pp. 102230
Author(s):  
Shuchao Pang ◽  
Matthew Field ◽  
Jason Dowling ◽  
Shalini Vinod ◽  
Lois Holloway ◽  
...  
PLoS ONE ◽  
2018 ◽  
Vol 13 (11) ◽  
pp. e0207001 ◽  
Author(s):  
Kang-Yi Su ◽  
Jeng-Sen Tseng ◽  
Keng-Mao Liao ◽  
Tsung-Ying Yang ◽  
Kun-Chieh Chen ◽  
...  

2016 ◽  
Vol 27 (2) ◽  
pp. 336-351 ◽  
Author(s):  
Akram Shalabi ◽  
Masato Inoue ◽  
Johnathan Watkins ◽  
Emanuele De Rinaldis ◽  
Anthony CC Coolen

When data exhibit imbalance between a large number d of covariates and a small number n of samples, clinical outcome prediction is impaired by overfitting and prohibitive computation demands. Here we study two simple Bayesian prediction protocols that can be applied to data of any dimension and any number of outcome classes. Calculating Bayesian integrals and optimal hyperparameters analytically leaves only a small number of numerical integrations, and CPU demands scale as O(nd). We compare their performance on synthetic and genomic data to the mclustDA method of Fraley and Raftery. For small d they perform as well as mclustDA or better. For d = 10,000 or more mclustDA breaks down computationally, while the Bayesian methods remain efficient. This allows us to explore phenomena typical of classification in high-dimensional spaces, such as overfitting and the reduced discriminative effectiveness of signatures compared to intra-class variability.


2015 ◽  
Vol 61 (1) ◽  
pp. 227-242 ◽  
Author(s):  
Arman Rahmim ◽  
C Ross Schmidtlein ◽  
Andrew Jackson ◽  
Sara Sheikhbahaei ◽  
Charles Marcus ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document