Stable Feature Selection from Brain sMRI
This work addresses the need for reproducible feature selection in neuroimaging for Alzheimer's disease diagnosis, representing an incremental improvement over existing sparse models.
The paper tackled the problem of unstable feature selection in brain sMRI for Alzheimer's disease diagnosis by proposing a nonnegative generalized fused lasso model that incorporates spatial cohesion and positive correlation priors, resulting in improved stable feature selection compared to state-of-the-art methods.
Neuroimage analysis usually involves learning thousands or even millions of variables using only a limited number of samples. In this regard, sparse models, e.g. the lasso, are applied to select the optimal features and achieve high diagnosis accuracy. The lasso, however, usually results in independent unstable features. Stability, a manifest of reproducibility of statistical results subject to reasonable perturbations to data and the model, is an important focus in statistics, especially in the analysis of high dimensional data. In this paper, we explore a nonnegative generalized fused lasso model for stable feature selection in the diagnosis of Alzheimer's disease. In addition to sparsity, our model incorporates two important pathological priors: the spatial cohesion of lesion voxels and the positive correlation between the features and the disease labels. To optimize the model, we propose an efficient algorithm by proving a novel link between total variation and fast network flow algorithms via conic duality. Experiments show that the proposed nonnegative model performs much better in exploring the intrinsic structure of data via selecting stable features compared with other state-of-the-arts.