In this paper, a new fuzzy clustering, namely fuzzy c-weighted mean (FCWM), is being proposed. The cost function of the classical fuzzy c-mean (FCM) is defined by the distances from data to the cluster centers with their fuzzy memberships. Another idea for estimating the cluster centers originating form the idea of weighted mean applied in nonparametric weighted feature extraction (NWFE) is introduced to established a novel FCM-like clustering algorithm in this study. The real data experimental results show that the proposing FCWM outperforms the original FCM.
31 October 2011 in Expert Systems with Applications
The well known fuzzy partition clustering algorithms are most based on Euclidean distance function, which can only be used to detect spherical structural clusters. Gustafson-Kessel (GK) clustering algorithm and Gath-Geva (GG) clustering algorithm, were developed to detect non-spherical structural clusters, but both of them based on semi-supervised Mahalanobis distance needed additional prior information. An improved fuzzy C-mean algorithm based on unsupervised Mahalanobis distance, FCM-M, was proposed by our previous work, but it didn't consider the relationships between cluster centers in the objective function. In this paper, we proposed an improved fuzzy C-mean algorithm, FCM-MS, which is not only based on unsupervised Mahalanobis distance, but also considering the relationships between cluster centers, and the relationships between the center of all points and the cluster centers in the objective function, the singular and the initial values problems were also solved. A real data set was applied to prove that the performance of the FCM-MS algorithm gave more accurate clustering results than the FCM and FCM-M methods, and the ratio method which is proposed by us is the better of the two methods for selecting the initial values.
Fuzzy clustering has been used widely in education, statistics, engineering, communication... etc. The well known fuzzy possibility c-mean algorithm can improve the problems of outlier and noise in fuzzy c-mean, but it was based on Euclidean distance function, which can only be used to detect spherical structural clusters. Extending Euclidean distance to Mahalanobis distance, Gustafson-Kessel (GK) clustering algorithm and Gath-Geva (GG) clustering algorithm were developed to detect non-spherical structural clusters, but these two algorithms fail to consider the relationships between cluster centers in the objective function. Yin-Tang-Sun-Sun (YTSS) clustering algorithm solved the relationships between cluster centers question, unfortunately, they did not consider the distance between the center of all data points and the center of each cluster. This problem was solved and presented in this paper. In this paper, a new fuzzy clustering algorithm (FPCM-S) was developed based on the conventional fuzzy c-means (FCM) to obtain more accurate clustering results with new separable criterion. It is different from YTSS cluster algorithm. The improved equations for the membership and the cluster center were derived from the alternating optimization algorithm. The noise and outlier were considered to obtain more accurate clustering results. Numerical data showed that the FPCM-S clustering algorithm gave more accurate clustering results than those of both FCM and YTSS clustering algorithms.
In this paper, a new fuzzy density function, called N-density, is proposed. A real data set about Students Valuing Science with 5-fold cross-validation RMSE is conducted, for comparing the performances of the Choquet integral regression model with respect to six measures, P-measure, λ-measure, L-measure, extensional L-measure, completed L-measure and extensional completed L-measure based on this new density and the old fuzzy support fuzzy density function, R-density, respectively, and two traditional regression model, multiple regression model and ridge regression model, the results show that the Choquet integral regression model with respect to each given fuzzy measure based on the new fuzzy density function is better than which based on the old fuzzy density function, and among all of the regression models, Choquet integral with respect to extensional completed L-measure based on N-density has the best performance.