Sedimentary Environment Analysis by Grain-Size Data Based on Mini Batch K-Means Algorithm

Algorithm 1

Pseudocode of Mini Batch K-means algorithm for sediment sample data processing.

Mini Batch K-Means

Input: the dataset of grain size is X; the number of initial clusters k is 3; the iteration times is t;

   the mini batch is b.

Output: the set of clustering labels is C; the cluster label of every sample is c.

Initialize every sample label as .

;

for i = 1 to t do

              //extract randomly mini batch sub-samples from .

   for do

     ;    //calculate and storage the clustering central sample

   end for

   for do

     ;       //acquire the central sample

     ;    //update the per-center counter

     ;      //get the real-time per-center learning rates

     ; //take gradient step

   end for

end for