The two main reasons to perform conditioned Latin hypercube (clhc) to recommend soil sample locations are (1) to ensure a good spread across your training data for models and (2) to obtain the optimal locations for a limited amount of sample to be taken.
The problem arises when historic soil samples exist and the agro wants to take additional samples. Instead of disregarding the historic soil samples, they should be taken into account when running clhc for a new set of samples; such that near-duplicate samples are not produced.
Solution:
The solution proposed, is to run the objective functions on the historic sample locations + the set of proposed sample locations; then to choose a set of proposed sample locations that, with the historic sample locations produce the best total objective factor.
Problem:
The two main reasons to perform conditioned Latin hypercube (clhc) to recommend soil sample locations are (1) to ensure a good spread across your training data for models and (2) to obtain the optimal locations for a limited amount of sample to be taken.
The problem arises when historic soil samples exist and the agro wants to take additional samples. Instead of disregarding the historic soil samples, they should be taken into account when running clhc for a new set of samples; such that near-duplicate samples are not produced.
Solution:
The solution proposed, is to run the objective functions on the historic sample locations + the set of proposed sample locations; then to choose a set of proposed sample locations that, with the historic sample locations produce the best total objective factor.