Citation: Jinliang Wang, Hongying Shu. Global analysis on a class of multi-group SEIR model with latency and relapse[J]. Mathematical Biosciences and Engineering, 2016, 13(1): 209-225. doi: 10.3934/mbe.2016.13.209
[1] | Gideon Simpson, Daniel Watkins . Relative entropy minimization over Hilbert spaces via Robbins-Monro. AIMS Mathematics, 2019, 4(3): 359-383. doi: 10.3934/math.2019.3.359 |
[2] | Kai Wang, Xiheng Wang, Xiaoping Wang . Curvature estimation for point cloud 2-manifolds based on the heat kernel. AIMS Mathematics, 2024, 9(11): 32491-32513. doi: 10.3934/math.20241557 |
[3] | Guangshuai Zhou, Chuancun Yin . Family of extended mean mixtures of multivariate normal distributions: Properties, inference and applications. AIMS Mathematics, 2022, 7(7): 12390-12414. doi: 10.3934/math.2022688 |
[4] | Pingyun Li, Chuancun Yin . Tail risk measures with application for mixtures of elliptical distributions. AIMS Mathematics, 2022, 7(5): 8802-8821. doi: 10.3934/math.2022491 |
[5] | Gaosheng Liu, Yang Bai . Statistical inference in functional semiparametric spatial autoregressive model. AIMS Mathematics, 2021, 6(10): 10890-10906. doi: 10.3934/math.2021633 |
[6] | Carey Caginalp . A minimization approach to conservation laws with random initialconditions and non-smooth, non-strictly convex flux. AIMS Mathematics, 2018, 3(1): 148-182. doi: 10.3934/Math.2018.1.148 |
[7] | H. M. Barakat, M. H. Dwes . Asymptotic behavior of ordered random variables in mixture of two Gaussian sequences with random index. AIMS Mathematics, 2022, 7(10): 19306-19324. doi: 10.3934/math.20221060 |
[8] | Stefano Bonaccorsi, Bernard Hanzon, Giulia Lombardi . A generalized Budan-Fourier approach to generalized Gaussian and exponential mixtures. AIMS Mathematics, 2024, 9(10): 26499-26537. doi: 10.3934/math.20241290 |
[9] | Miyoun Jung . A variational image denoising model under mixed Cauchy and Gaussian noise. AIMS Mathematics, 2022, 7(11): 19696-19726. doi: 10.3934/math.20221080 |
[10] | Essam A. Ahmed, Laila A. Al-Essa . Inference of stress-strength reliability based on adaptive progressive type-Ⅱ censing from Chen distribution with application to carbon fiber data. AIMS Mathematics, 2024, 9(8): 20482-20515. doi: 10.3934/math.2024996 |
Cao et al. [4] reviewed techniques for extracting local features for automatic object recognition in images. Multivariate Gaussians can represent the important features and their mutual correlations needed for accurate document retrieval from databases. The natural choice for discrimination between pairs of such distributions is the Fisher information metric on the Riemannian manifold of smooth probability density functions coordinatized by the parameters of the distribution [1,2]. However, it is not known analytically in some important cases of practical interest.
We have used multivariate Gaussians for face recognition using the neighbourhoods of colour pixel features at landmark points in face images [12], where we found that the spatial covariances among pixel colours was important. Craciunesco and Murari et al [5,8] used geodesic distance on Gaussian manifolds to interpret time series in very large databases from Tokomak measurements in fusion research. Verdoolaege, Shabbir et al [10,13] used multivariate generalized Gaussians for colour texture discrimination in the wavelet domain. In these studies the discrimination used approximations to the information distance between pairs of multivariate Gaussian probability density functions. Nielsen et al [9] suggested an entropic quantization method for approximating distances in the case of mixtures of multivariate Gaussians.
The
f(μ,Σ)=e−12(x−μ)TΣ−1(x−μ)√(2π)k|Σ|, | (1) |
where
The Riemannian manifold
Here we have the positive definite symmetric quadratic form
Dμ(fA,fB)=√(μA−μB)T⋅Σ−1⋅(μA−μB). | (2) |
So also we have a norm on mean vectors for each
||μA||=√(μA)T⋅(ΣA)−1⋅(μA) | (3) |
which is evidently sensitive to the covariance.
Here we use a positive definite symmetric matrix constructed from
SAB=ΣA−1/2⋅ΣB⋅ΣA−1/2, with {λABj}=Eig(SAB) then |
DΣ(fA,fB)=√12k∑j=1log2(λABj). | (4) |
We note that (4) is in agreement also with a special case of the geodesic distance given by Shabbir et al [10] for generalized multivariate Gaussians with the same mean.
In principle, (4) yields all of the geodesic distances since the information metric is invariant under affine transformations of the mean [3] Appendix 1; see also the article of P.S. Eriksen [6]. The equations for the geodesics were shown by Skovgaard [11] to be
¨μ=˙ΣΣ−1˙μ¨Σ=˙ΣΣ−1˙Σ−˙μ˙μT. | (5) |
Eriksen [6] observed that the family
˙Δ=−BΔ+vδT with Δ(0)=I,δ(0)=0,˙δ=−Bδ+(1+δTΔ−1δ)v. | (6) |
Then using
A=(−Bv0vT0−vT0−vB) | (7) |
Eriksen proved that the geodesic solution curve is given by
Λ:R:→Mk:t↦eAt=(ΔδΦδT1+δTΔ−1δγTΦTγΓ) | (8) |
where γ=Δ−1δ+ΦTΔ−1δ, and δTΔ−1δ = γTΓ−1γ. | (9) |
So (Δ(−t),δ(−t)) = (Γ(t),δ(t)). | (10) |
Of course, the analytic difficulty is the requirement to find the length of the geodesic between two points in
Here we consider a mixture distribution consisting of a linear combination of
f2(μ2,Σ2),f3(μ3,Σ3)...,fN(μN,ΣN) and ∀k∫Rkfk=1 | (11) |
where
Σ=k∑i≤j=1σijEij. |
We presume that the parameters and relative weights
fA=N∑k=2wAkfAk, with wAk≥0 and N∑k=2wAk=1. | (12) |
Given two such distributions,
There is no general analytic solution for the geodesic distance between two
So, a fortiori, also we do not have the distance between two mixtures of multivariate Gaussians:
Perhaps the most natural method is to combine equations (2) and (4) through the linear combination (12), obtaining an approximation as a corresponding linear combination of distances. Given two mixture distributions
D#μ(fA,fB)=N∑k=212(wAk√δμT⋅ΣA−1k⋅δμ+wBk√δμT⋅ΣBk−1⋅δμ) | (13) |
D#Σ(fA,fB)=N∑k=212(wAk√12N∑k=2(logλABk)2+wBk√12N∑k=2(logλBAk)2) {λABk}=EigHABk, HABk=((ΣAk)−1/2⋅ΣBk⋅(ΣAk)−1/2). | (14) |
We note
| | | |
Mixture | 0.6085 | 0.5635 | 0.5868 |
Mixture | 0.7087 | 0.7694 | 0.7522 |
Mixture | 0.9002 | 0.7126 | 0.7774 |
| | | |
Mixture | 0.8607 | 0.8110 | 0.8537 |
Mixture | 0.8607 | 0.8110 | 0.8537 |
Mixture | 0.8607 | 0.8110 | 0.8537 |
As expected from the form of
However, the
The new implementation described here uses the information geometric norm on the mean vectors and the Frobenius norm on the covariance matrices to project the mixture distributions onto the complex plane. This 2-dimensional representation reveals influences of the means and covariances in the mixtures, which itself may be a valuable It allows also the direct calculation of a distance between two mixture distributions using moduli, without assuming any connections between the mixtures, though this has the effect of smoothing the component influences of the means and covariances.
The idea here is simple: for each mixture distribution
||Mαβ||2=n∑α=1n∑β=1(mαβ)2 |
Note that if
||Mαβ||2=n∑α=1(λα)2. |
Given a mixture distribution
fA=M∑m=1wAmGAm |
||μA||=√M∑m=1wAm(μAm)T.(ΣAm)−1.(μAm) | (15) |
||ΣA||=√M∑m=1wAm||ΣAm||2. | (16) |
Now we can represent
Δ(fA,fB)=|ϕB−ϕA|. | (17) |
Figure 2 shows a plot of the points
There are efficient software programs for extracting from large data sets and image sequences certain mixtures of probability distributions, such as multivariate Gaussians, to represent the important features and their mutual correlations needed for accurate document retrieval from databases. The lack of an analytic solution to the geodesic distance equations between points in the Riemannian space of multivariate Gaussian mixtures, Equation (12), with an information metric, means that approximate solutions need to be found for practical applications. We have illustrated a new approximation for the case of 250 mixtures of
Whereas there are not analytic expressions for the information geometric distance between pairs of mixtures of multivariate Gaussians, we have shown that there are several choices for good information geometric approximate distances which are easy to compute. The new method yielded evident discrimination between pairs of these mixtures, shown in easily interpretable graphical form, Figure 2, distinguishing effects of covariance changes and effects of weighting sequences.
The methods described in §2.1 were developed with J. Scharcanski and J. Soldera during a visit to UFRGS, Brazil with a grant from The London Mathematical Society in 2013 and the author is grateful to CAPES (Coordeanao de Aperfeioamento de Pessoal de Nivel Superior, Brazil) for partially funding this project. An application of the results to face recognition was reported elsewhere in a joint paper [12].
The author declares that there are no conflicts of interest in this paper.
[1] | in: L.I. Lutwick (eds.), Tuberculosis: A Clinical Handbook, Chapman and Hall, London, 1995, 54-101. |
[2] | Nature, 280 (1979), 361-367. |
[3] | Oxford University Press, Oxford, 1991. |
[4] | Funkcial. Ekvac., 31 (1988), 331-347. |
[5] | in: T.G. Hallam, L.J. Gross, S.A. Levin (eds.), Mathematical Ecology, World Scientific, Teaneck NJ, (1988), 317-342. |
[6] | Academic Press, New York, 1979. |
[7] | in: Lecture Notes in Mathematics, Vol. 35, Springer, Berlin, 1967. |
[8] | American Public Health Association, Washington, 1999. |
[9] | J. Theor. Biol., 291 (2011), 56-64. |
[10] | J. Math. Biol., 28 (1990), 365-382. |
[11] | Clin. Inf. Dis., 33 (2001), 1762-1769. |
[12] | J. Differential Equations, 218 (2005), 292-324. |
[13] | Canad. Appl. Math. Quart., 14 (2006), 259-284. |
[14] | Proc. Amer. Math. Soc., 136 (2008), 2793-2802. |
[15] | Appl. Math. Sci., vol. 99, Springer, New York, 1993. |
[16] | Tran. R. Soc. Trop. Med. Hyg., 94 (2000), 247-249. |
[17] | Theor. Popu. Biol., 14 (1978), 338-349. |
[18] | in: Lecture Notes in Mathematics, vol. 1473, Springer-Verlag, berlin, 1991. |
[19] | SIAM J. Appl. Math., 52 (1992), 835-854. |
[20] | Bull. Math. Biol., 71 (2009), 75-83. |
[21] | Math. Biosci., 28 (1976), 221-236. |
[22] | Regional Conference Series in Applied Mathematics, SIAM, Philadelphia, 1976. |
[23] | SIAM J. Appl. Math., 70 (2010), 2434-2448. |
[24] | J. Differential Equations, 248 (2010), 1-20. |
[25] | J. Math. Anal. Appl., 361 (2010), 38-47. |
[26] | Nonlinear Anal.: RWA, 12 (2011), 119-127. |
[27] | National Academy Press, Washington, 1994. |
[28] | W.A. Benjamin Inc., New York, 1971. |
[29] | Nonlinear Anal.: RWA, 13 (2012), 1581-1592. |
[30] | Appl. Math. Comput., 218 (2011), 280-286. |
[31] | in: Mathematics in Biology and Medicine, Lecture Notes in Biomathematics, Springer, 57 (1995), 185-189. |
[32] | Math. Biosci. Eng., 4 (2007), 205-219. |
[33] | Math. Biosci., 207 (2007), 89-103. |
[34] | J. Amer. Med. Assoc., 259 (1988), 1051-1053. |
[35] | J. Biol. Dyn., 8 (2014), 99-116. |
[36] | J. Biol. Syst., 20 (2012), 235-258. |
[37] | Math. Comput. Simul., 79 (2008), 500-510. |
| | | |
Mixture | 0.6085 | 0.5635 | 0.5868 |
Mixture | 0.7087 | 0.7694 | 0.7522 |
Mixture | 0.9002 | 0.7126 | 0.7774 |
| | | |
Mixture | 0.8607 | 0.8110 | 0.8537 |
Mixture | 0.8607 | 0.8110 | 0.8537 |
Mixture | 0.8607 | 0.8110 | 0.8537 |