Improved K-Means Algorithm and Its Application to Vehicle Steering Identification

K-means is a very common clustering algorithm, whose performance depends largely on the initially selected cluster center. The K-means algorithm proposed by this paper uses a new strategy to select the initial cluster center. It works by calculating the minimum and maximum distances from data to the origin, dividing this range into several equal ranges, and then adjusting every range according to the data distribution to equate the number of data contained in the ranges as much as possible, and finally calculating the average of data in every range and taking it as initial cluster center. The theoretical analysis shows that despite linear time complexity of initialization process, this algorithm has the features of an superlinear initialization method. The application of this algorithm to the analysis of GPS data when vehicle is moving shows that it can effectively increase the clustering speed and finally achieve better vehicle steering identification.

[1]  Sergei Vassilvitskii,et al.  Scalable K-Means by ranked retrieval , 2014, WSDM.

[2]  M. Goyal,et al.  Improving the Initial Centroids of k-means Clustering Algorithm to Generalize its Applicability , 2014 .

[3]  Mothd Belal Al-Daoud.  A New Algorithm for Cluster Initialization , 2005, WEC.

[4]  Stephen J. Redmond,et al.  A method for initialising the K-means clustering algorithm using kd-trees , 2007, Pattern Recognit. Lett..

[5]  Sergei Vassilvitskii,et al.  k-means++: the advantages of careful seeding , 2007, SODA '07.

[6]  M. P. Sebastian,et al.  Improving the Accuracy and Efficiency of the k-means Clustering Algorithm , 2009 .

[7]  Meena Mahajan,et al.  The planar k-means problem is NP-hard , 2012, Theor. Comput. Sci..

[8]  Pierre Hansen,et al.  NP-hardness of Euclidean sum-of-squares clustering , 2008, Machine Learning.

[9]  Mohammad Al Hasan,et al.  Robust partitional clustering by outlier and density insensitive seeding , 2009, Pattern Recognit. Lett..

[10]  Patricio A. Vela,et al.  A Comparative Study of Efficient Initialization Methods for the K-Means Clustering Algorithm , 2012, Expert Syst. Appl..

[11]  Yanheng Liu,et al.  GPS-Based Vehicle Moving State Recognition Method and Its Applications on Dynamic In-Car Navigation Systems , 2014, 2014 IEEE 12th International Conference on Dependable, Autonomic and Secure Computing.

[12]  M. Emre Celebi,et al.  Improving the performance of k-means for color quantization , 2011, Image Vis. Comput..

[13]  José Antonio Lozano,et al.  An efficient approximation to the K-means clustering for massive data , 2017, Knowl. Based Syst..