10-04-2017, 09:37 PM
Implementation of rough fuzzy k-means clustering algorithm in Matlab
With the assistance of the lower and upper approximation of rough sets, the rough fuzzy k-means clustering algorithm may improve the objective function and further the distribution of membership function for the traditional fuzzy k-means clustering. However, the algorithm only has theoretical ideas rather than concrete realizations. To make it better applied to practice, using Matlab, a mathematical programming tool, to implement rough fuzzy k-means clustering algorithm is discussed. Moreover, steps of implementation are given in detail. The foresaid contributions may provide clustering learners and non-computer professional researchers with a simple, convenient, efficient and feasible implementation method.
clustering using fuzzy rough set feature selection
abstract
clustering also a form of data grouping, groups a set of data such that intra cluster similarity is maximized and the inter cluster similarity is minimized.
using feature selection method unimportant or irrelevant features are eliminated, and smaller set of attributes are generated.
frfs(fuzzy rough feature selection) and antfrfs is used for feature selection.
selected features are clustered using sdr(standard deviation roughness) algorithm.
performance is compared.
rough set theory
rough set theory can be regarded as a new mathematical tool for imperfect data analysis.
the theory has found applications in many domains, such as decision support, engineering, environment, banking, medicine and others.
objective
the main objective is to compare the accuracy of two feature selection algorithm namely frfs(fuzzy rough feature selection) and antfrfs algorithm in terms of clustering.
problem statement
selecting features using two algorithms, frfs and antfrfs separately and comparing their accuracy in clustering using sdr clustering.
conclusion
the study states clustering using fuzzy rough feature selection or antfrfs will give best accuracy, depending on the future work.