Abstract:In k-means clustering algorithm, it remains a problem that the
initial cluster centers are selected randomly. To deal with the
randomness problem, the Prim minimum spanning tree is introduced to
the initial center selection of K-means clustering algorithm.
Firstly, the prim method is used to find the minimum spanning tree
for the randomly generated points, then a group of points are
selected as the initial center which has the maximum sum of the
weights and all weights have little difference. Finally, we propose
an improved K-means clustering algorithm based on prim. Simulation
results show that, compared with NKM (Normal K-Means) algorithm, our
algorithm improves the accuracy and reduces the data iterations.
Kun Wang;Lili Xu;Yan Li;Zhixin Sun;Meng Wu;Zhen Yang. IKCAP: An Improved K-means Clustering Algorithm Based on Prim[J]. , 2013, 10(13): 4303-4310.
Kun Wang;Lili Xu;Yan Li;Zhixin Sun;Meng Wu;Zhen Yang. IKCAP: An Improved K-means Clustering Algorithm Based on Prim. , 2013, 10(13): 4303-4310.