Method of clusterization of quasiperiodic signal based on clonal selection algorithm
Abstract
The paper proposes a hierarchical-iterative method of clustering of a quasi-periodic signal based on clonal selection algorithm, which increases the speed and accuracy of clustering. Preliminary transformation of samples (quasi-periodic areas) of this signal into a single amplitude-time window based on shift and scaling in time and amplitude, linear interpolation and time sampling is performed. Signals processed by intelligent computer systems for identification, analysis and synthesis, storage of digital signals (such as acoustic, graphic, vibration, electrogram, communication ones, etc.) are often quasi-periodic. This raises the problem of constructing effective methods for analyzing the structure of a quasi-periodic signal. One of the means of analyzing the structure of a quasiperiodic signal is clustering, which is a type of machine learning without a teacher. As a result of the analysis of modern methods of clustering of quasi-periodic signals, it has been found that most of them have one or more of the following disadvantages: unknown exact number of clusters; sensitivity to initial values of the centroids of clusters; low probability of clustering; low speed of clustering; comparison of signal areas that have only the same size; comparison of only binary signals. Therefore, it is important to develop a method for transforming a quasi-periodic signal and a hierarchical-iterative method of clustering based on a clonal selection algorithm. This will increase the efficiency for the analysis of quasi-periodic signal structure in digital data processing in intelligent computer systems of identity identification, technical and medical diagnostics, network traffic analysis, etc. A comparison of the method proposed by the authors and existing clustering methods is given, with the cloning parameter α = 0.1, the mutation parameter β = 2.5, the number of replacement antibodies d = 0.2 | H |. Clustering has been performed on quasi-periodic speech sounds uttered by different speakers. A method for converting a quasi-periodic signal, which converts samples (quasiperiodic areas) of this signal into a single amplitude-time window by shifting and scaling in time and amplitude, interpolation and sampling, is proposed. This allows to compare signal samples of different lengths and with different amplitudes. A method for clustering a quasi-periodic signal based on a hierarchical-iterative approach and a clonal selection algorithm has been developed, which reduces the sensitivity to initial values of cluster centroids due to random search and provides adaptive adjustment of the cluster number due to the hierarchical approach, and also increases the probability of clustering to 0.98
Keywords
quasi-periodic signal; clonal selection algorithm; hierarchical-iterative clustering; signal transformation; signal structure analysis
References
[1] K. S. Thyagarajan, Introduction to Digital Signal Processing using MATLAB with Application to Digital Communications. Cham, Switzerland: Springer, 2020.
[2] P. R. Hill, Audio and Speech Processing with MATLAB. Boca Raton, FL, USA: CRC Press, 2019.
[3] D. Barber, Bayesian Reasoning and Machine Learning. Cambridge, GB: Cambridge University Press, 2013.
[4] G. James, D. Witten, T. Hastie, and R. Tibshirani, An Introduction to Statistical Learning with Applications in R. New York, USA: Springer, 2013. doi: 10.1007/978-1-4614-7138-7.
[5] T. Hastie, R. Tibshiriani, and M. Wainwright, Statistical Learning with Sparsity. The Lasso and Generalizations. New York, USA: Chapman and Hall/CRC, 2015. doi: 10.1201/b18401.
[6] A. N. Tkachenko, O. F. Griyo Tukalo, O. V. Dzisy, and S. M. Lakhovets, "Clustering method based on sequential launch of k-means with improved selection of a candidate for a new insertion position", Naukovi pratsi VNTU, no. 2, pp. 1-10, 2012 [in Russian].
[7] M. J. Brusco, E. Shireman, and D. Steinley, "A comparison of latent class, k-means, and k-median methods for clustering dichotomous data", Psychological Methods, 22 (3), pp. 563-580, 2017. doi: 10.1037/met0000095.
[8] C. C. Aggarwal, and C. K. Reddy, Data Clustering: Algorithms and Applications. Boca Raton, FL, USA: CRC Press, 2014.
[9] Z. Fu, and L. Wang, "Color image segmentation using Gaussian mixture model and EM algorithm", in Proc. Multimedia and Signal Processing. Communications in Computer and Information Science (CMSP 2012), 2012, vol. 346, pp. 61-66.doi: 10.1007/978-3-642-35286-7_9
[10] T. Kohonen, Self-Organizing and Associative Memory, 3rd ed. Berlin, Heidelberg, Germany: Springer, 2012. doi: 10.1007/978-3-642-88163-3
[11] T. Kohonen, "Essentials of the selforganizing map", Neural Networks: The Official Journal of the Inter-National Neural Network Society, vol. 37, pp. 52-65, 2013. doi: 10.1016/j.neunet.2012.09.018.
[12] K.-L. Du, and M. N. S. Swamy, Neural Networks and Statistical Learning. London, GB: Springer-Verlag, 2014. doi: 10.1007/978-1-4471-5571-3
[13] Yu. A. Skobtsov, and E. E. Fedorov, Metaheuristics. Donetsk, Ukraine: Noulidzh (Donetsk branch), 2013 [in Russian].
[14] J. Radosavljević, Metaheuristic Optimization in Power Engineering. New York, USA: Institution of Engineering and Technology, 2018. doi: 10.1049/pbpo131e.
[15] E. Alba, A. Nakib, and P. Siarry, Metaheuristics for Dynamic Optimization. Berlin, Germany: Springer-Verlag, 2013. doi: 10.1007/978-3-642-30665-5.
[16] K.-L. Du, and M. N. S. Swamy, Search and Optimization by Metaheuristics. Techniques and Algorithms Inspired by Nature. Charm, Switzerland: Springer, 2016. doi: 10.1007/978-3-319-41192-7.
[17] O. Bozorg Haddad, M. Solgi, and H. Loaiciga, Meta-heuristic and Evolutionary Algorithms for Engineering Optimization. Hoboken, New Jersey, USA: Wiley & Sons, 2017. doi: 10.1002/9781119387053.
[18] X.-S. Yang, Nature-Inspired Algorithms and Applied Optimization. Charm, Switzerland: Springer, 2018. doi: 10.1007/978-3-319-67669-2.
[19] L. N. De Castro, and F. J. Von Zuben, "Learning and optimization using the clonal selection principle", IEEE Transactions on Evolutionary Computation, Special Issue on Artificial Immune Systems, vol. 6, no. 3, pp. 239-251, 2002.
[20] E. D. Ülker, and S. Ülker, "Comparison study for clonal selection algorithm and genetic algorithm", International Journal of Computer Science & Information Technology (IJCSIT), vol. 4, no. 4, pp. 107-118, 2012. doi: 10.5121/ijcsit.2012.4410