A feature selection algorithm should ideally satisfy four conditions: reliably extract relevant features; be able to identify non-linear feature interactions; scale linearly with the number of features and dimensions; allow the incorporation of known sparsity structure. In this work we propose a novel feature selection algorithm, Gradient Boosted Feature Selection (GBFS), which satisfies all four of these requirements. The algorithm is flexible, scalable, and surprisingly straight-forward to implement as it is based on a modification of Gradient Boosted Trees. We evaluate GBFS on several real world data sets and show that it matches or outperforms other state of the art feature selection algorithms. Yet it scales to larger data set sizes and naturally allows for domain-specific side information.

No abstract

Submarine groundwater discharge (SGD) is a global phenomenon that carries large volumes of groundwater and dissolved chemical species such as nutrient, metals, and organic compounds to coastal zones. We report the influence of SGD on the coastal waters of Jeju Island, Korea, using high‐resolution aerial thermal infrared (TIR) mapping techniques and field investigations. An aircraft‐based system was implemented using a cost‐effective TIR camera for aerial TIR mapping. Ground‐based calibrations and system integration with GPS/IMU (global positioning system/inertial measurement unit) were performed for the aerial systems. The aerial surveys showed distinct low‐temperature signatures of SGD along the coasts of Jeju Island, revealing large groundwater inputs from the coastal aquifers to the ocean. Multiple aerial surveys over a range of seasons and tidal stages revealed that SGD rates dynamically affect the sea surface temperature (SST) of the coastal zone. The in‐situ measurements supported that SGD has a substantial influence on the coastal water chemistry as well as SST. Our observations highlight the extent to which aerial‐based TIR mapping can serve as a powerful tool for studying SGD and other coastal processes. Copyright © 2016 John Wiley & Sons, Ltd.

In order to select proper seed points for triangle constrained image-matching propagation, this letter analyzes the affects of different numbers and different distributions of seed points on the image-matching results. The concept of distribution quality is introduced to quantify the distribution of seed points. An intensive experimental analysis is illustrated using two different stereo aerial images and, based on the experimental results, a seed point selection strategy for triangle constrained image-matching propagation is proposed. An automatic selection method is then introduced that gives good distribution quality for a defined number of seed points.