2019 VOL.2 Apr No.2

• Title: Research on functional data clustering method with constraint condition
• Name: Yi-min Cui
• Company: School Of Information Engineering, Xi'an University, Xi'an Shaan
• Abstract:

  Most existing clustering methods only cluster from distance, and do not make full use of information contained in functional data, and the form of functional data does not conform to the nature of distribution function. To address these problems, a functional data clustering method with constraint condition is proposed in this paper. A stochastic process framework is used to define the function type data, and the constraint problem is transformed into unconstrained case. The physical background of the energy smoothing method is analyzed. The curve is regarded as an elastic spline, and the strain energy of the elastic spline is used as the energy function of the curve. The function data curve is smoothed through the energy optimization method. Combining the characteristics of functional data, clustering is based on the actual distance of functional data. The further clustering of data with similar intrinsic characteristics in each class is achieved through the distance of the derived function. The distance of functional data is measured by the principal component distance between functions. The intrinsic feature similarity of data is measured by the distance of the derived function. Experiment is carried out for the proposed method applied to the clustering of the Doppler signal, and the results show that the proposed method can achieve effective classification.

• Keyword: Constraint condition; function type; data; clustering; smoothing
• DOI: 10.12250/jpciams2019020119
• Citation form: Yi-min Cui.Research on functional data clustering method with constraint condition[J]. Computer Informatization and Mechanical System, 2019, vol. 2, pp. 36-41.

[1] Abin,A.A., Beigy, H.(2015). Active constrained fuzzy clustering: A multiple kernels learning approach. Pattern Recognition, 48(3):953-967.

[2] Antoine, V., Quost, B., Masson, M. H., et al.(2014).CEVCLUS: evidential clustering with instance-level constraints for relational data. Soft Computing, 18(7):1321-1335.

[3]Bakhtiarifar, M. H., Bashiri, M.(2015). A probabilistic clustering method for data elements with normal distributed attributes. Communications in Statistics - Simulation and Computation, 46(4):2563-2575.

[4]Deng, F.(2017). Multiple Hops Network Classification Attribute Data Fuzzy Clustering in the Simulation. Computer Simulation, 34(1):292-295.

[5] Liu, X., Li, M.(2014). Integrated constraint based clustering algorithm for high dimensional data. Neurocomputing,142(1):478-485.

[6] Kou, G., Peng, Y., Wang, G.(2014). Evaluation of clustering algorithms for financial risk analysis using MCDM methods. Information Sciences,275(11):1-12.