In the literature, finite mixture models were described as linear combinations of probability distribution functions having the form $ f(x) = \Lambda \sum\limits_{i = 1}^n w_i f_i(x) $, $ x \in \mathbb{R} $, where $ w_i $ were positive weights, $ \Lambda $ was a suitable normalising constant, and $ f_i(x) $ were given probability density functions. The fact that $ f(x) $ is a probability density function followed naturally in this setting. Our question was: if we removed the sign condition on the coefficients $ w_i $, how could we ensure that the resulting function was a probability density function?
The solution that we proposed employed an algorithm which allowed us to determine all zero-crossings of the function $ f(x) $. Consequently, we determined, for any specified set of weights, whether the resulting function possesses no such zero-crossings, thus confirming its status as a probability density function.
In this paper, we constructed such an algorithm which was based on the definition of a suitable sequence of functions and that we called a generalized Budan-Fourier sequence; furthermore, we offered theoretical insights into the functioning of the algorithm and illustrated its efficacy through various examples and applications. Special emphasis was placed on generalized Gaussian mixture densities.
Citation: Stefano Bonaccorsi, Bernard Hanzon, Giulia Lombardi. A generalized Budan-Fourier approach to generalized Gaussian and exponential mixtures[J]. AIMS Mathematics, 2024, 9(10): 26499-26537. doi: 10.3934/math.20241290
In the literature, finite mixture models were described as linear combinations of probability distribution functions having the form $ f(x) = \Lambda \sum\limits_{i = 1}^n w_i f_i(x) $, $ x \in \mathbb{R} $, where $ w_i $ were positive weights, $ \Lambda $ was a suitable normalising constant, and $ f_i(x) $ were given probability density functions. The fact that $ f(x) $ is a probability density function followed naturally in this setting. Our question was: if we removed the sign condition on the coefficients $ w_i $, how could we ensure that the resulting function was a probability density function?
The solution that we proposed employed an algorithm which allowed us to determine all zero-crossings of the function $ f(x) $. Consequently, we determined, for any specified set of weights, whether the resulting function possesses no such zero-crossings, thus confirming its status as a probability density function.
In this paper, we constructed such an algorithm which was based on the definition of a suitable sequence of functions and that we called a generalized Budan-Fourier sequence; furthermore, we offered theoretical insights into the functioning of the algorithm and illustrated its efficacy through various examples and applications. Special emphasis was placed on generalized Gaussian mixture densities.
| [1] | H. Albrecher, D. Finger, P. O. Goffard}, Empirical risk analysis of mining a proof of work blockchain, 2023. Available from: https://www.researchgate.net/publication/375082859 |
| [2] | L. Ambrosio, N. Gigli, G. Savaré, Gradient flows in metric spaces and in the space of probability measures, Lectures in Mathematics ETH Zürich, Birkhäuser Verlag, Basel, (2008). |
| [3] |
L. Aquilanti, S. Cacace, F. Camilli, R. De Maio, A mean field games approach to cluster analysis, Appl. Math. Optim., 84 (2021), 299–323. https://doi.org/10.1007/s00245-019-09646-2 doi: 10.1007/s00245-019-09646-2
|
| [4] |
L. Aquilanti, S. Cacace, F. Camilli, R. De Maio, A mean field games model for finite mixtures of Bernoulli and categorical distributions, J. Dyn. Games, 8 (2021), 35–59. https://doi.org/10.3934/jdg.2020033 doi: 10.3934/jdg.2020033
|
| [5] | S. Bonaccorsi, B. Hanzon, G. Lombardi, GBF-algorithm, 2024. Available from: https://github.com/giuliaelle/GBF-algorithm |
| [6] |
V. Chonev, J. Ouaknine, J. Worrell, On the zeros of exponential polynomials, J. ACM, 70 (2023), 1–26. https://doi.org/10.1145/3603543 doi: 10.1145/3603543
|
| [7] |
P. Collins, Computable analysis with applications to dynamical systems, Math. Structures Comput. Sci., 30 (2020), 173–233. https://doi.org/10.1017/S096012952000002X doi: 10.1017/S096012952000002X
|
| [8] |
M. Coste, T. Lajous-Loaeza, H. Lombardi, M. F. Roy, Generalized Budan-Fourier theorem and virtual roots, J. Complexity, 21 (2005), 479–486. https://doi.org/https://doi.org/10.1016/j.jco.2004.11.003 doi: 10.1016/j.jco.2004.11.003
|
| [9] | G. Dall'Aglio, Sugli estremi dei momenti delle funzioni di ripartizione doppia, Ann. Scuola Norm. Sup. Pisa Cl. Sci., 10 (1956), 35–74. |
| [10] | T. A. Driscoll, N. Hale, L. N. Trefethen, eds., Chebfun Guide, Pafnuty Publications, Oxford, (2014). |
| [11] | R. M. Dudley, Probabilities and metrics, Matematisk Institut, Aarhus Universitet, Aarhus, (1976). |
| [12] |
P. Embrechts and M. Hofert, A note on generalized inverses, Math. Methods Oper. Res., 77 (2013), 423–432. https://doi.org/10.1007/s00186-013-0436-7 doi: 10.1007/s00186-013-0436-7
|
| [13] | B. S. Everitt, S. Landau, M. Leese, D. Stahl, Cluster analysis, Wiley Series in Probability and Statistics, John Wiley & Sons, Ltd., Chichester, (2011), fifth eds. https://doi.org/10.1002/9780470977811 |
| [14] | S. Frühwirth-Schnatter, Finite mixture and Markov switching models, Springer Series in Statistics, Springer, New York, (2006). |
| [15] |
A. Galligo, Budan tables of real univariate polynomials, J. Symb. Computat., 53 (2013), 64–80. https://doi.org/https://doi.org/10.1016/j.jsc.2012.11.004 doi: 10.1016/j.jsc.2012.11.004
|
| [16] |
L. Gonzalez-Vega, H. Lombardi, L. Mahé, Virtual roots of real polynomials, J. Pure Appl. Algebra, 124 (1998), 147–166. https://doi.org/10.1016/S0022-4049(96)00102-8 doi: 10.1016/S0022-4049(96)00102-8
|
| [17] | B. Hanzon, F. Holland, Non-negativity analysis for exponential-polynomial-trigonometric functions on $[0, \infty)$, in Spectral theory, mathematical system theory, evolution equations, differential and difference equations, (eds. W. Arendt, J. Ball, J. Behrndt, K.H. Foerster, V. Mehrmann, C. Trunk), vol. 221 of Oper. Theory Adv. Appl., Birkhäuser/Springer Basel AG, Basel, (2012), 399–412. https://doi.org/10.1007/978-3-0348-0297-0_21 |
| [18] |
B. Hanzon, R. J. Ober, A state-space calculus for rational probability density functions and applications to non-gaussian filtering, SIAM J. Control Optim., 40 (2001), 724–740. https://doi.org/10.1137/S036301299731610X doi: 10.1137/S036301299731610X
|
| [19] |
R. Jiang, M. Zuo, H. X. Li, Weibull and inverse Weibull mixture models allowing negative weights, Reliab. Eng. Syst. Safe., 66 (1999), 227–234. https://doi.org/https://doi.org/10.1016/S0951-8320(99)00037-X doi: 10.1016/S0951-8320(99)00037-X
|
| [20] | A. Karapiperi, M. Redivo-Zaglia, M. R. Russo, Generalizations of Sylvester's determinantal identity, (2015). Available from: https://arXiv.org/abs/1503.00519 |
| [21] | S. Kasyanov, GaussElimination, (2022). Available from: https://www.mathworks.com/matlabcentral/fileexchange/90661-gausselimination |
| [22] |
W. Koch, On exploiting 'negative' sensor evidence for target tracking and sensor data fusion, Inform. Fusion, 8 (2007), 28–39. https://doi.org/https://doi.org/10.1016/j.inffus.2005.09.002 doi: 10.1016/j.inffus.2005.09.002
|
| [23] |
T. Li, H. Liang, B. Xiao, Q. Pan, Y. He, Finite mixture modeling in time series: A survey of Bayesian filters and fusion approaches, Inform. Fusion, 98 (2023), 101827. https://doi.org/10.1016/j.inffus.2023.101827 doi: 10.1016/j.inffus.2023.101827
|
| [24] | G. McLachlan, D. Peel, Finite mixture models, Wiley Series in Probability and Statistics: Applied Probability and Statistics, Wiley-Interscience, New York, (2000). https://doi.org/10.1002/0471721182 |
| [25] | P. Müller, S. Ali-Löytty, M. Dashti, H. Nurminen, R. Piché, Gaussian mixture filter allowing negative weights and its application to positioning using signal strength measurements, in 9th Workshop on Positioning, Navigation and Communication, Dresden, Germany, (2012), 71–76. https://doi.org/10.1109/WPNC.2012.6268741 |
| [26] | J. Park, I. W. Sandberg, Universal approximation using radial-basis-function networks, 3 (1991), 246–257. https://doi.org/10.1162/neco.1991.3.2.246 |
| [27] | K. Pearson, Contribution to the mathematical theory of evolution, Philos. T. A, 185 (1894), 71–110. |
| [28] |
G. Pólya, On the mean-value theorem corresponding to a given linear homogeneous differential equation, Trans. Amer. Math. Soc., 24 (1922), 312–324. https://doi.org/10.2307/1988819 doi: 10.2307/1988819
|
| [29] |
C. Ridders, A new algorithm for computing a single root of a real continuous function, IEEE T. Circuits Syst., 26 (1979), 979–980. https://doi.org/10.1109/TCS.1979.1084580 doi: 10.1109/TCS.1979.1084580
|
| [30] |
R. Ristroph, Pólya's property $W$ and factorization—A short proof, Proc. Amer. Math. Soc., 31 (1972), 631–632. https://doi.org/10.2307/2037591 doi: 10.2307/2037591
|
| [31] | D. H. H. Santosh, P. Venkatesh, P. Poornesh, L. N. Rao, N. A. Kumar, Tracking multiple moving objects using gaussian mixture model, Int. J. Soft Comput. Eng., 3 (2013), 114–119. |
| [32] |
C. Sexton, M. Olivi, B. Hanzon, Rational approximation of transfer functions of nonnegative ept densities, IFAC Proceedings, 45 (2012), 716–721. https://doi.org/10.3182/20120711-3-BE-2027.00197 doi: 10.3182/20120711-3-BE-2027.00197
|
| [33] | C. Sexton, Financial modelling with 2-EPT probability density functions, PhD Thesis, University College Cork, (2013). https://cora.ucc.ie/handle/10468/1430 |
| [34] | D. M. Titterington, A. F. M. Smith, U. E. Makov, Statistical analysis of finite mixture distributions, Wiley Series in Probability and Mathematical Statistics: Applied Probability and Statistics, John Wiley & Sons, Ltd., Chichester, (1985). |
| [35] | L. N. Trefethen, Approximation Theory and Approximation Practice, Society for Industrial and Applied Mathematics, (2019). |
| [36] | S. S. Vallander, Calculations of the Vasseršteĭn distance between probability distributions on the line, Teor. Probab. Appl., 18 (1973), 824–827. |
| [37] | K. Weihrach, Computable Analysis, Texts in Theoretical Computer Science, Springer, (2000). https://doi.org/10.1007/978-3-642-56999-9 |
| [38] | D. Yu, L. Deng, Automatic speech recognition, Signals and Communication Technology, Springer, London, (2015). |
| [39] | B. Zhang, C. Zhang, Finite mixture models with negative components, in Machine Learning and Data Mining in Pattern Recognition, (eds. P. Perner and A. Imiya), Springer, Berlin, Heidelberg, (2005), 31–41. https://doi.org/10.1007/11510888_4 |
Figures(8) / Tables(1)
Stefano Bonaccorsi, Bernard Hanzon, Giulia Lombardi. A generalized Budan-Fourier approach to generalized Gaussian and exponential mixtures[J]. AIMS Mathematics, 2024, 9(10): 26499-26537. doi: 10.3934/math.20241290
DownLoad: