Learning from scanners: Bias reduction and feature correction in radiomics

Radiomics and Deep Learning

Radiomics in Brain Tumor: Image Assessment, Quantitative Feature Descriptors, and Machine-Learning Approaches.

Radiomics describes a broad set of computational methods that extract quantitative features from radiographic images. The resulting features can be used to inform imaging diagnosis, prognosis, and therapy response in oncology. However, major challenges remain for methodologic developments to optimize feature extraction and provide rapid information flow in clinical settings. Equally important, ...

Feature Reduction for Unsupervised Learning

In this project, four unsupervised feature reduction algorithms for clustering problem were investigated and experimented upon two sets of data – handwritten digits data set and the functional magnetic resonance imaging (fMRI) resting state data set. Ratio of sum of squares (RSS), leverage score (LEV), and Laplacian score (LAP) were used to rank the influences of the features in the clustering....

Linkage and Autocorrelation Cause Feature Selection Bias in Relational Learning

Two common characteristics of relational data sets — concentrated linkage and relational autocorrelation — can cause learning algorithms to be strongly biased toward certain features, irrespective of their predictive power. We identify these characteristics, define quantitative measures of their severity, and explain how they produce this bias. We show how linkage and autocorrelation affect a r...

Incremental Active Learning with Bias Reduction

The problem of designing input signals for optimal generalization in supervised learning is called active learning. In many active learning methods devised so far, the bias of the learning results is assumed to be zero. In this paper, we remove this assumption and propose a new active learning method with the bias reduction. The effectiveness of the proposed method is demonstrated through compu...