Interval kernel Fuzzy C-Means clustering of incomplete data

Li, Tianhao; Zhang, Liyong; Lu, Wei; Hou, Hui; Liu, Xiaodong; Pedrycz, Witold; Zhong, Chongquan

doi:10.1016/j.neucom.2017.01.017

Cited by 51 publications

(14 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The membership matrix is U = [u ik ], and Although FCM has wide applicability in various domains [2,30,47], it suffers from the issue of random initialization of the cluster centers and the tendency of its cost function to be stuck in a local optima [7]. To overcome such limitations, several extensions of the traditional FCM such as intuitionistic fuzzy set [46], picture fuzzy set [42] and kernel fuzzy set [29,37] have been proposed. However, intuitionistic FCM takes more number of iterations to find out the number of cluster centers than FCM, resulting in high computational time [46].…”

Section: Fuzzy C-meansmentioning

confidence: 99%

“…Similarly, in case of picture fuzzy set, an extra exponent parameter value is required to be set to obtain best fuzzy cluster sets, thus requiring more computational time [42]. Likewise, for kernel-based FCM, the problem lies in selecting the best kernel to find out the optimal distance of each point from the cluster center, which is a quite tedious process [29,37]. Hence, we have chosen the classical FCM algorithm in the current work rather than its variants and applied GA on it for optimizing the cluster centers by searching a global optimum to make the clustering approach more robust.…”

Section: Fuzzy C-meansmentioning

confidence: 99%

See 1 more Smart Citation

A hybrid mobile call fraud detection model using optimized fuzzy C-means clustering and group method of data handling-based network

Subudhi

Panigrahi

2018

Vietnam J Comput Sci

View full text Add to dashboard Cite

A novel two-stage fraud detection system in mobile telecom networks has been presented in this paper that identifies the malicious calls among the normal ones in two stages. Initially, a genetic algorithm-based optimized fuzzy c-means clustering is applied to the user's historical call records for constructing the calling profile. Thereafter, the identification of the fraudulent calls occurs in two stages. In the first stage, each incoming call is passed to the clustering module that identifies the call as genuine, malicious or suspicious. This is done by comparing the distance value of the new calling instance from the profile cluster centers against two predefined threshold values. The calls detected as genuine or malicious are not further processed. However, the call records that are found to be suspicious are additionally scrutinized in the second stage by a previously trained group method of data handling model for final decision making. The legitimate and forged labeled call records generated out of the clustering module are utilized for training the supervised classifier. Experimentation is done on a real-world call dataset to exhibit the effectiveness of the proposed model. A comparative analysis of the current approach with one of our earlier propositions and another recent fraud detection system clearly illustrates the efficacy of the developed model. Keywords Fraud detection • Mobile phone call records • Genetic algorithm • Fuzzy C-means • Group method of data handling

show abstract

Section: Fuzzy C-meansmentioning

confidence: 99%

Section: Fuzzy C-meansmentioning

confidence: 99%

A hybrid mobile call fraud detection model using optimized fuzzy C-means clustering and group method of data handling-based network

Subudhi

Panigrahi

2018

Vietnam J Comput Sci

View full text Add to dashboard Cite

show abstract

“…These features, especially incompleteness, lead to the widespread use of incomplete data in practical applications [ 15 , 16 ]. Lack of data values will affect the decision process of the application servers for specific tasks [ 17 ]. The resulting errors can be important for subsequent steps in data processing.…”

Section: Introductionmentioning

confidence: 99%

The Optimally Designed Variational Autoencoder Networks for Clustering and Recovery of Incomplete Multimedia Data

Xiang

Zhang

et al. 2019

Sensors

View full text Add to dashboard Cite

Clustering analysis of massive data in wireless multimedia sensor networks (WMSN) has become a hot topic. However, most data clustering algorithms have difficulty in obtaining latent nonlinear correlations of data features, resulting in a low clustering accuracy. In addition, it is difficult to extract features from missing or corrupted data, so incomplete data are widely used in practical work. In this paper, the optimally designed variational autoencoder networks is proposed for extracting features of incomplete data and using high-order fuzzy c-means algorithm (HOFCM) to improve cluster performance of incomplete data. Specifically, the feature extraction model is improved by using variational autoencoder to learn the feature of incomplete data. To capture nonlinear correlations in different heterogeneous data patterns, tensor based fuzzy c-means algorithm is used to cluster low-dimensional features. The tensor distance is used as the distance measure to capture the unknown correlations of data as much as possible. Finally, in the case that the clustering results are obtained, the missing data can be restored by using the low-dimensional features. Experiments on real datasets show that the proposed algorithm not only can improve the clustering performance of incomplete data effectively, but also can fill in missing features and get better data reconstruction results.

show abstract

“…If for a datum a i the values of some features are missing, then the datum a i is said to be incomplete, and the whole set A is called an incomplete data set (see e.g. [15,20]).…”

Section: Introductionmentioning

confidence: 99%

“…Methods for clustering data with missing values are also considered in the Master's thesis by [38]. In the paper [20], missing values are estimated in the form of intervals using the nearest neighbor method. Several papers about image analysis in case of data missing in an image or a partial missing edge image can be found in the proceedings [1].…”

Section: Introductionmentioning

confidence: 99%

Searching for an Optimal Partition of Incomplete Data with Application in Modeling Energy Efficiency of Public Buildings

2018

View full text Add to dashboard Cite

In this paper, we consider the problem of searching for an optimal partition with the most appropriate number of clusters for an incomplete data set A ⊂ R n in which several outliers might occur. Special attention is given to the application of the Least Squares distance-like function. The procedure of preparing the incomplete data set and the outlier elimination procedure are proposed such that the clustering process gives acceptable solutions. Appropriate justifications with proof are provided for these procedures. An incremental algorithm for searching for optimal partitions with 2, 3,. .. clusters is applied on the prepared data set. After that, by using the Davies-Bouldin and the Calinski-Harabasz index the most appropriate number of clusters is determined. The whole procedure is organized as an algorithm given in the paper. In order to illustrate its applicability, the above steps are applied on the real data set of public buildings and their energy efficiency data, providing clear clusters that could be used for further modeling procedures.

show abstract

Interval kernel Fuzzy C-Means clustering of incomplete data

Cited by 51 publications

References 34 publications

A hybrid mobile call fraud detection model using optimized fuzzy C-means clustering and group method of data handling-based network

A hybrid mobile call fraud detection model using optimized fuzzy C-means clustering and group method of data handling-based network

The Optimally Designed Variational Autoencoder Networks for Clustering and Recovery of Incomplete Multimedia Data

Searching for an Optimal Partition of Incomplete Data with Application in Modeling Energy Efficiency of Public Buildings

Contact Info

Product

Resources

About