Free Access

Fig. 3.


Fits for linear basis (left panel) and polynomial basis (right panel) for random forests and LASSO. Note that the horizontal lines seen in the fits are there because of decision tree regression: decision trees set one value of the y-axis for a given range of x-axis leading to a step-function like fit. In both cases, LASSO seems to do a better job at capturing the curved shape of the vertical profiles. The black vertical dashed line presents the train-test split (80% training data).

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.