We are doing the prediction of soil nutrients using spectral information by applying machine learning models. More than 300 soil samples were used from one district for analysis. We got the support vector machine model as best with less RMSE. Is it the same devolved model that holds good for other district soil nutrient prediction?
I assume that you have a model for predicting soil nutrients from spectral signatures (I assume) and this model is based on SVM. It can hold as long as it is trained on similar spectral signatures and on recognizing similar soil nutrients based on the same parameters. Otherwise, you have to train your model again to predict newly introduced spectral signatures, newly introduced nutrients, or changed parameters.
if you change the number classes by introducing spectral signatures and nutrients, you have to train it again. Otherwise for same classes you can make slight changes in hyper parameters and tune your model
Thank you, Shanaka Ramesh. Please clarify for me about the change in the number of classes.
Same nutrients we are evaluating but soil condition is different.
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Soil Science
- Soil Analysis