Seagrass detection in the mediterranean: A supervised learning approach

Dimitrios Effrosynidis, Avi Arampatzis, Georgios Sylaios

Ecological Informatics · 2018 · 49 citations · 42 references

DOIFull text

Open access

Concepts

Abstract

We deal with the problem of detecting seagrass presence/absence and distinguishing seagrass families in the\nMediterranean via supervised learning methods. By merging datasets about seagrass presence and other external\nenvironmental variables, we develop suitable training data, enhanced by seagrass absence data algorithmically\nproduced based on certain hypotheses. Experiments comparing several popular classification algorithms yield up\nto 93.4% accuracy in detecting seagrass presence. In a feature strength analysis, the most important variables\ndetermining presence–absence are found to be Chlorophyll-α levels and Distance-to-Coast. For determining family,\nvariables cannot be easily singled out; several different variables seem to be of importance, with\nChlorophyll-α surpassing all others. In both problems, tree-based classification algorithms perform better than\nothers, with Random Forest being the most effective. Hidden preferences reveal that Cymodocea and Posidonia\nfavor the low, limited-range chlorophyll-α levels (< 0.5 mg/m3), Halophila tolerates higher salinities (> 39),\nwhile Ruppia prefers euryhaline conditions (37.5–39).

References

42