Exploring the similarity of medical imaging classification problems
This work addresses the need for efficient method selection in medical image analysis, though it is incremental as it focuses on similarity quantification rather than full meta-learning.
The paper tackles the problem of quantifying similarity between medical imaging classification problems as a first step towards meta-learning, achieving 89.3% accuracy in classifying datasets by origin using a meta-feature space based on classifier performance ranks.
Supervised learning is ubiquitous in medical image analysis. In this paper we consider the problem of meta-learning -- predicting which methods will perform well in an unseen classification problem, given previous experience with other classification problems. We investigate the first step of such an approach: how to quantify the similarity of different classification problems. We characterize datasets sampled from six classification problems by performance ranks of simple classifiers, and define the similarity by the inverse of Euclidean distance in this meta-feature space. We visualize the similarities in a 2D space, where meaningful clusters start to emerge, and show that the proposed representation can be used to classify datasets according to their origin with 89.3\% accuracy. These findings, together with the observations of recent trends in machine learning, suggest that meta-learning could be a valuable tool for the medical imaging community.