Performance comparison of machine learning algorithms for land cover classification
Description
This dataset consists of shapefiles of training points names (SamplePoints.zip) which were collected on the Google Earth Engine platform for our land cover analysis. The code.zip contains the codes and the direct Google Earth Engine code showing how the land cover analysis with six (6) machine learning algorithms (Random Forest (RF), Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), Gradient Tree Boost (GTB), Classification and Regression Tree (CART), and Naive Bayes) was implemented. Each link to a code has been annotated with the correct season (either fall, spring, summer, and winter). Additionally, these links have been added as related links to this dataset strictly in the order Fall > Spring > Summer > Winter.
Files
Steps to reproduce
To reproduce the works, 1.User/Reproducer should have a Google Earth Engine account. 2. Open the code for the season you want to reproduce. A. Open the first link to reproduce for the FALL season. B. Open the second link to reproduce for the SPRING season. C. Open the third link to reproduce for the SUMMER season. D. Open the fourth link to reproduce for the WINTER season. 3. To run the code for a specific algorithm, go to the algorithm section of the code and comment out all unwanted algorithms, making sure only the algorithm of interest is active or uncommented. 4. To download the data for a particular season and algorithm, change the filename at the export data section and run the code
Institutions
- Kwame Nkrumah University of Science and Technology Department of Geomatic Engineering
- University of Maine
- Westfalische Wilhelms-Universitat Munster Institut fur Geoinformatik
- Mississippi State University
- University of South Florida