The algorithm proposed by Ziko et.al normalizes the data that is read from csv, possibly to optimise calculations. In this process the data, that was originally [latitude, longitude] loses meaning. Therefore the obtained set of final cluster centers output meaningless data.
Note that as the algorithm being used is Kmeans, the final set of cluster centers may not be actual points hence they cannot be directly denormalized by multiplying by their L2 norms (the L2 norms for these points may and probably don't exist in the first place).
To do
Attempts need to be made to denormalize the cluster centers so that the final output is [latitude, longitude] for each cluster centers
Description
The algorithm proposed by Ziko et.al normalizes the data that is read from csv, possibly to optimise calculations. In this process the data, that was originally [latitude, longitude] loses meaning. Therefore the obtained set of final cluster centers output meaningless data. Note that as the algorithm being used is Kmeans, the final set of cluster centers may not be actual points hence they cannot be directly denormalized by multiplying by their L2 norms (the L2 norms for these points may and probably don't exist in the first place).
To do
Attempts need to be made to denormalize the cluster centers so that the final output is [latitude, longitude] for each cluster centers
Example
Initial points = [ [0.335051 1.273698] [0.396364 1.548297] [0.386358 1.545602] [0.495796 1.358735] [0.452601 1.606864] [0.187534 1.61443 ] [0.226379 1.354459] [0.322582 1.289061] [0.405518 1.533587] [0.333909 1.272195] ]
No. of clusters = 3
Final (normalized )cluster centers = [ [-0.86672733 0.08333477] [ 0.62955802 0.5882229 ] [-0.21662397 -0.97434153] ]
Final (denormalized) cluster centers = [???]
Resources that might help