<p>Boxplots of the core Canberra distance between lists of selected features obtained using different methods when 50, 20, 15 or 10 subjects per group are available. A star highlights the significant differences between pair of bootstrap and non-bootstrap approaches (p-value lower than 0.05, Wilcoxon test).</p
Feature selection plays an important role in applications with high dimensional data. The assessment...
Results of bootstrap tests for the difference in components of segment shape variation between popul...
<p>The data was simulated according to <i>x</i><sub>1<i>i</i></sub> = <i>ϵ</i><sub>1<i>i</i></sub>, ...
<p>Boxplots of precision corresponding to the optimal number of features chosen by different methods...
The AIC difference between two Poisson-linked models of cluster growth for all four full data sets, ...
For each architecture, a set of deep feature lists is generated, one list for each internal run of t...
<p>Boxplots of area under the precision <i>vs.</i> recall curves obtained by ranking features accord...
<p>The stabilities (as relative ranks from 0.0 = most unstable to 1.0 = most stable) of the shared f...
<p>The features (abbreviated names are transparently based on the full WALS names and as for <a href...
<p>The partial lists are either identical, randomly permuted (average distance) or maximally distant...
It is well known that variable selection in multiple regression can be unstable and that the model u...
Within each growth layout, a logistic regression was performed to classify connection existence from...
<p>Stability is measured as the percent of voxels in common among the subsets of <i>k</i> top variab...
Results of bootstrap tests for the difference in components of segment shape variation between popul...
<p>Bootstrap values, that is, the percentage of trees yielding the same two-set partition of sequenc...
Feature selection plays an important role in applications with high dimensional data. The assessment...
Results of bootstrap tests for the difference in components of segment shape variation between popul...
<p>The data was simulated according to <i>x</i><sub>1<i>i</i></sub> = <i>ϵ</i><sub>1<i>i</i></sub>, ...
<p>Boxplots of precision corresponding to the optimal number of features chosen by different methods...
The AIC difference between two Poisson-linked models of cluster growth for all four full data sets, ...
For each architecture, a set of deep feature lists is generated, one list for each internal run of t...
<p>Boxplots of area under the precision <i>vs.</i> recall curves obtained by ranking features accord...
<p>The stabilities (as relative ranks from 0.0 = most unstable to 1.0 = most stable) of the shared f...
<p>The features (abbreviated names are transparently based on the full WALS names and as for <a href...
<p>The partial lists are either identical, randomly permuted (average distance) or maximally distant...
It is well known that variable selection in multiple regression can be unstable and that the model u...
Within each growth layout, a logistic regression was performed to classify connection existence from...
<p>Stability is measured as the percent of voxels in common among the subsets of <i>k</i> top variab...
Results of bootstrap tests for the difference in components of segment shape variation between popul...
<p>Bootstrap values, that is, the percentage of trees yielding the same two-set partition of sequenc...
Feature selection plays an important role in applications with high dimensional data. The assessment...
Results of bootstrap tests for the difference in components of segment shape variation between popul...
<p>The data was simulated according to <i>x</i><sub>1<i>i</i></sub> = <i>ϵ</i><sub>1<i>i</i></sub>, ...