Inland water systems are essential to our environment because they are vital ecosystems that are bio-diverse. Thus, finding innovative ways to monitor water quality is vital. In this repository, I present various machine learning algorithms that takes in multispectral remote sensing data from the AquaSat data set as input to predict optically active water quality parameters as desired output.
In this experiment, several water quality parameters were to be predicted using only multispectral data. The parameters predicted in this repository include Chlorophyll-A, Dissolved Organic Carbon, Total Inorganic Sediment, and Total Suspended Sediment. Each water quality parameter was predicted separately using the same machine learning algorithms and multispectral data.
In this repository, a files is included for each water quality parameter to be predicted. Each file also shows the results of the predictions with an analysis of these predictions following. Some preprocessing and explanatory data analysis was also conducted in another file.