We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
USE OF MACHINE LEARNING ALGORITHMS AND IN SITU DATA FOR ESTIMATING PARTICULATE ORGANIC CARBON FROM THE MEDITERRANEAN SEA.
- Authors
S., FELLOUS; A., BENDJAMA; Y., BENZAOUI
- Abstract
Water quality indicators, including biological, chemical and physical properties, are usually determined by collecting data from the field and analyzing them in the laboratory. Although these in situ measurements are costly and time-consuming, they offer high accuracy. This study focuses on the estimation of particulate organic carbon (POC) as a water quality parameter using a combination of machine learning algorithms and hyperspectral in situ data. A data-driven approach that does not need any domain knowledge was used. We were interested in POC generated by bacteria, phytoplankton, zooplankton, detritus and sediments in the Mediterranean Sea from 15 May to 10 June 2017. Therefore, the objective of this study was to use five regression frameworks from machine learning algorithms to estimate POC with hyperspectral in situ data and evaluate their performance. Based on the coefficient of determination R², the best-performing modes were nearest neighbors (KNN), gradient boosting (GB) and random forest (RF), with an R² in the range of 72.33 to 74.7%. These machine learning models can be used to investigate more water quality parameters, as they reveal the great potential of this approach.
- Subjects
COLLOIDAL carbon; MACHINE learning; MACHINE theory; WATER quality; RANDOM forest algorithms; CHEMICAL properties
- Publication
Larhyss Journal, 2023, Issue 56, p179
- ISSN
1112-3680
- Publication type
Article