On the spherical cardioid distribution and its goodness-of-fit

Eduardo Garc\'ia-Portugu\'es

REVIEW 3 major objections 6 minor 1 cited by

On the spherical cardioid distribution and its goodness-of-fit

T0 review · 3 major / 6 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The spherical cardioid distribution is a fully tractable spherical model: closed under convolution, with explicit moments, estimators, and a bootstrap goodness-of-fit test that fits long-period comet orbital normals with k=2.

desk verdict Solid, honest development of the spherical cardioid as a working model; the distribution theory is competent, but the bootstrap GOF test lacks a validity proof and the application over-reads k=2. read the letter →

arxiv 2601.16095 v2 pith:IREVVUZZ submitted 2026-01-22 stat.ME math.STstat.TH

Eduardo García-Portugués This is my paper · ORCID

classification stat.MEmath.STstat.TH MSC 62H1162F0362F1233C45

keywords sphericalcardioiddistributiondirectionalstatisticsGegenbauerpolynomialsChebyshevmomentestimationmaximumlikelihoodgoodness-of-fitprojectedcumulativefunction

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper develops the spherical cardioid distribution, a family of distributions on the sphere defined as a uniform density plus a Gegenbauer-polynomial perturbation, and argues that it is a fully tractable statistical model. The paper establishes closedness under convolution, explicit vectorized moments (with all moments of order below the model order matching the uniform distribution), a closed-form characteristic function, efficient simulation algorithms, and consistent asymptotically normal moment and maximum-likelihood estimators. It then constructs a parametric-bootstrap goodness-of-fit test based on projected empirical distribution functions, with closed-form test statistics in low dimensions. Applied to orbital normals of long-period comets, the order-two spherical cardioid fits the data, while uniformity and orders one, three, and four are rejected. A sympathetic reader would care because this provides a simple, analytically workable alternative to the uniform distribution for mildly non-uniform spherical data.

What carries the argument

The key object is the normalized Gegenbauer polynomial $\tilde C_k^{(d-1)/2}(x^T \mu)$ (with Chebyshev polynomials for $d=1$), used as a perturbation of the uniform density on the sphere; its orthogonality and parity supply the moment-vanishing structure and the convolution formula. The estimators rely on the vectorized moments, and the goodness-of-fit test builds on the closed-form projected cumulative distribution function $F_\gamma$, which is the uniform projected cdf plus a $\rho$-times-polynomial term. The bootstrap in Algorithm 3 is the mechanism that turns the test statistic into p-values.

What would settle it

Simulate iid samples from $C_2(\mu,0.5)$ on $S^2$ with $n=100$, estimate $(\mu,\rho)$, run Algorithm 3 with $B=1000$, and check whether the resulting p-values are $\mathrm{Uniform}(0,1)$ under the null; substantial miscalibration would make the reported comet p-values uninterpretable.

Watch

Extended reading notes

Core claim

The central discovery is that the family $C_k(\mu,\rho)$ on $S^d$ — densities proportional to $1 + \rho \tilde C_k^{(d-1)/2}(x^T \mu)$, where $\tilde C$ is the normalized Gegenbauer (Chebyshev when $d=1$) polynomial — is closed under convolution in a simple sense, has vectorized moments computable in closed form, and admits moment and maximum-likelihood estimators with explicit asymptotic variances. A notable structural fact is that for any model of order $k$, all moments of order $m<k$ coincide with the uniform sphere's moments, as do moments of orders $m>k$ with $m-k$ odd; this makes high-order spherical cardioids nearly indistinguishable from uniformity by low-order moment information. The paper also derives the exa

Load-bearing premise

The goodness-of-fit test relies on the unproven assumption that the parametric bootstrap in Algorithm 3 yields valid p-values when the null parameters are estimated from the same sample; the paper states the procedure as standard and provides no theorem showing the bootstrap distribution approximates the null distribution.

Editorial extensions

If this is right

For k=1 and k=2, the method-of-moments and maximum-likelihood estimators are strongly consistent and asymptotically normal, with explicit variances; the asymptotic relative efficiency formulas show the moment estimator of concentration loses efficiency for large |ρ|.
The convolution closure means that mixing the location of one spherical cardioid by another spherical cardioid of the same order yields a spherical cardioid with a product-type concentration divided by a known dimension factor—useful for hierarchical modeling.
All moments of order < k match the uniform sphere, so a large-k spherical cardioid is a near-uniform alternative that is difficult for standard low-order uniformity tests to detect.
The projected cdf is explicit, so the paper's CvM and AD statistics have closed-form V-statistic expressions in d=1,2 for k=1,2, avoiding numerical integration in those cases.
For long-period comet orbital normals, the order-two cardioid is not rejected at the 10% level, whereas uniformity and orders 1,3,4 are rejected; for short-period comets every order is rejected.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A likely use of the moment-vanishing property is to construct high-order spherical cardioids as explicit alternatives for power studies of uniformity tests; the paper notes this as a challenge, but it is a direct consequence of its moment theorem.
The bootstrap test's validity under composite nulls is assumed rather than proved; if one filled that gap (e.g., via a conditional convergence argument or a corrected bootstrap), it would solidify the reported comet p-values.
The closed-form projected cdf could support other goodness-of-fit statistics (e.g., characteristic-function or energy distances) for the same family, an extension the paper discusses as future work.
The convolution closure may allow tractable mixtures or Bayesian hierarchical priors on S^d, since the location parameter stays within the same family.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Thanks for the report. I read the paper too. The core distribution-theoretic work is real: the moments, convolution closure, simulation algorithms, and estimator asymptotics are detailed, and I did not find a load-bearing error. The author also deserves credit for clearly citing Baringhaus–Grübel and Borodavka–Ebner; the novelty claim is modest but honest—nobody had treated this as a parametric model. The projected cdf and the closed-form CvM/AD kernels are a genuine service, even if the derivations lean on Mathematica.

The soft spots are where your report puts them. The parametric bootstrap in Algorithm 3 is the clear gap. The paper calls the procedure standard and gives no theorem or reference establishing that resampling from C_k(μ̂, ρ̂) yields valid p-values under the composite null. Table 1's size checks are only n=100 with B=100, and the conservative behaviour at ρ=0.75 shows the procedure is not obviously safe. I would want either a proof or a citation to a general bootstrap validity result for this kind of statistic before relying on the comet p-values.

The other issues are smaller. Theorem 5.3's kernel formulas are asserted after "Mathematica", with no derivations or code; that is hard to check and should be made reproducible. The Table 2 caption formula k0=(k+1) mod 1 is simply wrong, and the Section 6.1 bias statement about ρ is misworded—the bias in the standardized scale is not the actual bias. The comet analysis picks k=2 after looking at k=1,2,3,4, with no multiplicity correction; that is worth a sentence of caution, though not a reason to reject the application.

Net: the paper is a competent, useful addition to directional statistics. It should go to review. But the reviewer should push for a bootstrap validity argument or reference, and for reproducible derivations of the kernels. If those land, the GOF test becomes trustworthy. With the current gaps, I would treat the comet conclusions as suggestive rather than established.

Referee Report

3 major / 6 minor

Summary. The paper introduces the spherical cardioid distribution C_k(μ,ρ) on S^d, defined by density f_{C_k}(x;μ,ρ)=ω_d^{-1}{1+ρ\tilde C_k^{(d-1)/2}(x^⊤μ)}, as a higher-dimensional, higher-order generalization of the circular cardioid. It establishes tractability properties: rotational symmetry and various shape regimes (Sec. 3.1), closedness under convolution (Prop. 3.1), explicit vectorized moments whose low orders coincide with the uniform moments (Thm 3.1 and Cor. 3.1–3.2), characteristic/moment generating functions (Prop. 3.2), and two simulation algorithms (Algs 1–2). Estimation is treated by the method of moments for k=1,2 (Thms 4.1–4.2), a Gegenbauer-moment estimator for known location (Thm 4.3), and maximum likelihood with asymptotic normality (Thm 4.4), together with asymptotic relative efficiencies (Sec. 4.3). The second half develops a projected-ecdf goodness-of-fit test (Sec. 5): explicit projected density and cdf (Thm 5.1), V-statistic forms for general and Anderson–Darling weights (Thm 5.2, Cor. 5.1), closed-form kernels for the Cramér–von Mises statistic (Thm 5.3), and a parametric bootstrap procedure (Alg. 3). Numerical experiments and an application to long-period comet orbital normals suggest that C_2 provides an adequate model. The principal inferential claim for the goodness-of-fit test is the validity of the parametric bootstrap under estimated parameters, for which the paper provides no theorem or reference.

Significance. If the results hold, the spherical cardioid family is a genuinely useful addition to directional statistics: a simple one-concentration family on S^d that is close to uniformity, has closed-form density, moments, and characteristic function, and is easy to simulate. The moment formulas and the asymptotic distributions of the estimators are derived in detail, with explicit variance expressions; the ARE analysis is careful and informative. The projected-ecdf test extends a prior uniformity-testing framework to a parametric family and provides closed-form test statistics for several important cases. However, the paper's most novel inferential tool — the bootstrap goodness-of-fit test — is presented without a formal validity argument, and one of the closed-form results depends on unshown computer algebra. These gaps currently prevent full confidence in the reported comet-data p-values and in the claim that C_2 is an adequate model for long-period comet orbital normals. If the bootstrap validity is established (or the application is recast as exploratory), and the computational derivations are made verifiable, the paper would be a solid contribution to the field.

major comments (3)

[Sec. 5.3, Algorithm 3; Table 3] The parametric bootstrap is load-bearing for the goodness-of-fit conclusion. The statistic P_n^{W,λ} uses the projected cdf F̂_γ with parameters estimated from the same sample; its null distribution is therefore not the simple-hypothesis distribution. The paper states in Sec. 5.3 that the bootstrap procedure is 'standard' but gives no theorem, proof, or citation ensuring that bootstrap samples from C_k(μ̂,ρ̂) approximate the null distribution of P_n^{W,λ}. This is particularly important because the comet p-values in Table 3 — and the claim that C_2 is adequate for long-period comet orbital normals — rest on this procedure. The empirical size checks in Table 1 are limited to n=100 and a small grid of (ρ,k,d); several entries lie well outside the 95% prediction interval (e.g., k=1,d=2,ρ=0.75 rows, with rejection rates 1.4–2.6% against a 5% nominal level). These checks do not substitute for
[Sec. 5.2, Theorem 5.3; Appendix C.2] The closed-form kernels φ and ψ in Theorem 5.3 are a stated contribution and are used in the exact evaluation of P_n^{CvM,Unif}. The proof in Appendix C.2 delegates parts of the integral evaluations to Mathematica (e.g., the expressions for φ̃^(1) and φ̃^(2) in the proof of Theorem 5.3), without showing the symbolic derivations or providing an independently checkable verification. A referee cannot easily confirm that the formulas are correct, and a dependence on unshown computer algebra is especially delicate in a statistics paper where these formulas feed into test implementations and numerical experiments. Please either provide derivations (or at least a clear, verifiable reduction to standard integrals) or make the computer algebra notebook/code available, and state explicitly that the results have been independently checked.
[Sec. 6.1] The text states: 'The bias for μ_1 has order 10^{−2}, but for ρ the estimated bias ρ̄̂−ρ is still significant: approximately 0.20 for k=1 and 0.33 for k=2.' As written, this claims a bias of order 0.2–0.3 in ρ̂ at n=1000 with true ρ=0.5, which is inconsistent with the strong consistency and asymptotic normality established in Theorems 4.1–4.4. If these are biases of the standardized statistics √n(ρ̂−ρ), that should be stated explicitly and the values are plausible (they correspond to original-scale biases of order 10^{−2}). If they are not, the statement is erroneous and needs correction. This ambiguity undermines the numerical validation in Section 6.1.

minor comments (6)

[Sec. 6.2, Table 2 caption] The caption defines the null order as k_0=(k+1) mod 1, which is always 0. The text above the table says (k,k0) ∈ {(1,2),(2,1)}, so the formula is presumably a typo; please correct it.
[Sec. 6.2] The statement that the variable performance in Table 1 'can be explained by the fact that the Monte Carlo samples are shared within each row' is not a convincing explanation for the systematically low rejection rates in the (k=1,d=2,ρ=0.75) row. Shared random numbers explain correlation, not a level shift away from 5%. Please discuss or investigate whether this reflects conservativeness of the bootstrap test at n=100.
[Sec. 3.5] The sentence 'The inverse transformation method is rarely preferable over rejection sampling besides k=1,2' should read '...except for k=1,2'.
[Sec. 5.2] In the paragraph before Eq. (24), the notation for the Anderson–Darling statistic uses U_(i) but the definition of U_(i) is given later as ordered values of U_i^{(γ)}=F̂_γ(γ^⊤X_i). This is clear enough, but a parenthetical reminder would help.
[Figure 3] The y-axis is labeled 'ARE' without indicating which estimator (μ or ρ) is being plotted; the captions clarify, but the axis itself could be more informative.
[Sec. 4.3] The parenthetical '(see gray dashed lines)' in Sec. 6.1 is ambiguous because the gray dashed lines in Figure 5 represent the empirical mean of the standardized statistics, not a bias in the original scale; please align the text with the figure.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's derivations are self-contained up to standard orthogonal-polynomial identities and prior external lemmas; the Section 5.3 bootstrap gap is a validity concern, not a circular reduction.

full rationale

The spherical cardioid density is introduced by explicit definition (Definition 3.1), and the subsequent results are proven from that definition using Gegenbauer/Chebyshev orthogonal-polynomial identities, projection operators, and standard calculus. The moment theorems (e.g., Theorem 3.1) are direct consequences of the density's Fourier/Gegenbauer form, not independent predictions fitted to data. The estimators in Section 4 invert closed-form moments or score equations (e.g., Theorem 4.1 uses E[X] = ρ/(d+1)μ), so they are standard moment/MLE procedures and are never marketed as out-of-sample predictions. The goodness-of-fit test in Section 5 is constructed from the projected cdf with parameters estimated from the same sample; Algorithm 3's parametric bootstrap is asserted as 'standard' without a theorem (Section 5.3), which is a real omitted-proof/validity gap in the inferential claim, but it is not a circular reduction of the paper's equations. Citations to García-Portugués et al. (2023) are for orthogonal-polynomial and projected-ecdf lemmas that do not involve the spherical cardioid; they are auxiliary and parameter-free, so they do not make the central claims circular. No equation is shown to equal its own input by construction, and no fitted parameter is renamed as a prediction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The core derivations use standard orthogonal-polynomial machinery and the paper's own earlier projected-ecdf framework (García-Portugués et al. 2023). The main hidden inputs are: correctness of Mathematica-evaluated kernels, validity of the parametric bootstrap, and standard ML regularity conditions. The fitted comet parameters are not hidden ad hoc constants in the derivation, but they do support the application claim.

free parameters (3)

μ (location) for long-period comet fit = (0.0804, −0.0067, 0.9967)
ML estimate in Section 7; the application-level conclusion depends on this fitted vector, but the distribution theory does not.
ρ (concentration) for long-period comet fit = 0.4727
ML estimate in Section 7; the C_2 model fit depends on this value.
order k for comet model = 2 (chosen after testing k=1,2,3,4)
Model order selected post hoc from Table 3; no multiplicity correction. This affects the strength of the application claim.

assumptions (5)

standard math Completeness and orthogonality of Gegenbauer/Chebyshev polynomials on L^2_d([-1,1])
Section 2 invokes these to expand rotationally symmetric densities in zonal harmonics (Eq. 7); this is standard classical analysis.
standard math Uniform convergence of zonal expansions for smooth f via Kalf (1995, Thm 2)
Used to justify density expansions in Section 2; a standard smoothness result.
standard math ML regularity conditions (van der Vaart Theorems 5.41/5.42) hold for the ξ-parametrized likelihood
Invoked in the proof of Theorem 4.4 to obtain consistency and asymptotic normality of ML estimators in the interior parameter space.
ad hoc to paper The parametric bootstrap in Algorithm 3 yields valid p-values under estimated parameters
Section 5.3 states the bootstrap procedure without a validity theorem or a citation to a general bootstrap theorem; this is the weakest load-bearing assumption for the GOF contribution.
ad hoc to paper Mathematica evaluations in Theorem 5.3 of the φ and ψ kernels are correct
The text says the integrals were evaluated with Mathematica (Wolfram Research, Inc., 2021) but does not show the derivation; the test-statistic formulas depend on these computations.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the spherical cardioid distribution and its goodness-of-fit." pith.science (2026). https://pith.science/paper/IREVVUZZ

@misc{pith2026260116095,
  author       = {Pith},
  title        = {Pith review of: On the spherical cardioid distribution and its goodness-of-fit},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IREVVUZZ}},
  note         = {Machine review of arXiv:2601.16095}
}

abstract

In this paper, we study the spherical cardioid distribution, a higher-dimensional and arbitrary-order generalization of the circular cardioid distribution. This distribution is rotationally symmetric and generates unimodal, multimodal, axial, and girdle-like densities. We identify various properties of the spherical cardioid that make it highly tractable: simple density evaluation, closedness under convolution, explicit expressions for vectorized moments, and efficient simulation. The moments of the spherical cardioid of order $k$ up to order $k-1$ coincide with those of the uniform distribution on the sphere, highlighting its closeness to the latter. We derive estimators by the method of moments and maximum likelihood, their asymptotic distributions, and their asymptotic relative efficiencies. We give the machinery for bootstrap goodness-of-fit tests based on the projected empirical cumulative distribution function approach, including the projected distribution and closed-form expressions for test statistics. An application to modeling the orbits of long-period comets shows the usefulness of the spherical cardioid distribution in real data analyses.

Figures

Figures reproduced from arXiv: 2601.16095 by the authors.

**Figure 1.** Samples and density of the spherical cardioid on S 1 with µ = e2 (red point, north) and (k, ρ) ∈ {(1, 1),(2, 1),(2, −1),(3, 1),(4, 1),(4, −1)}. For each panel, a random sample of n = 200 observations is shown. The dashed curve gives the uniform density 1/(2π) as reference. Negative-ρ panels illustrate overparametrization. Just like the von Mises distribution with low concentration is approximately a circular cardio… view at source ↗

**Figure 2.** Samples and density of the spherical cardioid on S 2 with µ = e3 (red point, north) and (k, ρ) ∈ {(1, 1),(2, 1),(2, −1),(3, 1),(4, 1),(4, −1)}. The plots show the front hemisphere of S 2 , with shading applied to the points in the back hemisphere. The sample, with n = 2000 observations, is colored according to the value of the density at the observations. 3.2 Closedness under convolution The circular cardioid family… view at source ↗

**Figure 3.** shows the ARE curves ρ 7→ AREMM(µ) and ρ 7→ AREMM(ρ) for ρ ∈ (0, 1), k = 1, 2, and d = 1, . . . , 10. As expected, the AREs are smaller than one. The efficiency of both moment estimators is maximal for ρ = 0 and decreases as ρ increases, but differently for µˆMM and ρˆMM. Indeed, AREMM(µ) is almost always above 0.8, while AREMM(ρ) can attain minima below 0.25 for ρ ≈ 1. The AREs are larger for k = 2 than for k = 1 w… view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Asymptotic relative efficiencies ρ 7→ AREGM(ρ) for k = 3 and d = 1, . . . , 10 (left panel) and k = 1, . . . , 6 and d = 2 (right panel). 5 Goodness-of-fit We turn now our attention to the goodness-of-fit testing for the Ck(µ, ρ) distribution. Given an iid sample X1, .…

**Figure 5.** Figure 5: Histograms of { √ n(ˆµ (j) 1 −µ1)}M j=1 and { √ n(ˆρ (j) −ρ)}M j=1 against their asymptotic normal density, for k = 1, 2 and d = 2. Inside each panel, the left plot corresponds to the maximum likelihood estimator and the right plot corresponds to the method of moments …

**Figure 6.** Figure 6: Orbits of long- and short-period comets and their normal vectors. Within each figure, the left plot displays ten illustrative elliptical orbits and their associated normal vectors, with the ecliptic plane shown in gray and the Sun represented as an orange sphere (one f…

**Figure 7.** Figure 7: Comparison of the ecdf Fn,µˆML of the projected sample {µˆ ⊤ MLXi} n i=1 (black curve) versus the projected cdf FˆµˆML of the fitted Ck(µˆML, ρˆML) (red curve). 8 Discussion We have studied a generalization of the well-known circular cardioid distribution that allows f…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

M\"obius transport on spheres
math.ST 2026-07 accept novelty 7.0 of 10

Generalized Möbius transport—a cosine rearrangement fixed by any rotationally symmetric target—builds tractable anisotropic spherical distributions, including a new Möbius–von Mises–Fisher family.

Reference graph

Works this paper leans on

23 extracted references · 1 linked inside Pith · cited by 1 Pith paper

[1]

and Capitanio, A

Azzalini, A. and Capitanio, A. (2014).The Skew-Normal and Related Families, volume 3 ofInstitute of Mathematical Statistics Monographs. Cambridge University Press, Cambridge

2014
[2]

and Biswas, S

Banerjee, B. and Biswas, S. (2024). Intrinsic geometry-inspired dependent toroidal distribution: Application to regression model for astigmatism data.arXiv preprint, arXiv:2409.06229

arXiv 2024
[3]

and Grübel, R

Baringhaus, L. and Grübel, R. (2024). Discrete mixture representations of spherical distributions. Stat. Pap., 65(2):557–596

2024
[4]

and Ebner, B

Borodavka, J. and Ebner, B. (2026). A general maximal projection approach to uniformity testing on the hypersphere.Bernoulli, to appear. Chacón, J. E. and Duong, T. (2018).Multivariate Kernel Smoothing and its Applications, volume 160 ofMonographs on Statistics and Applied Probability. CRC Press, Boca Raton. Chacón, J. E., García-Portugués, E., and Meilán...

2026
[5]

and Xu, Y

Dai, F. and Xu, Y. (2013).Approximation Theory and Harmonic Analysis on Spheres and Balls. Springer Monographs in Mathematics. Springer, New York. 51 DLMF(2025).NIST Digital Library of Mathematical Functions. F.W.J.Olver, A.B.OldeDaalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller and B. V. Saunders, eds

2013
[6]

Dones, L., Brasser, R., Kaib, N., and Rickman, H. (2015). Origin and evolution of the cometary reservoirs.Space Sci. Rev., 197(1):191–269

2015
[7]

Ebner, B., Henze, N., and Meintanis, S. (2024). A unified approach to goodness-of-fit testing for spherical and hyperspherical data.Statistical Papers, 65(6):3447–3475. Fernández-de-Marcos, A. and García-Portugués, E. (2023). On new omnibus tests of uniformity on the hypersphere.Test, 32(4):1508–1529. Fernández-Durán, J. J. (2004). Circular distributions ...

2024
[8]

Gradshteyn, I. S. and Ryzhik, I. M. (2014).Table of Integrals, Series, and Products. Academic

2014

Show all 23 references

[9]

Jammalamadaka, S. R. and SenGupta, A. (2001).Topics in Circular Statistics, volume 5 ofSeries on Multivariate Analysis. World Scientific, Singapore

2001
[10]

(2003).Theory of Probability

Jeffreys, H. (2003).Theory of Probability. Oxford University Press, Oxford, third edition

2003
[11]

Jones, M. C. and Pewsey, A. (2005). A family of symmetric distributions on the circle.J. Am. Stat. Assoc., 100(472):1422–1428

2005
[12]

E., Kim, P

Jupp, P. E., Kim, P. T., Koo, J.-Y., and Wiegert, P. (2003). The intrinsic distribution and selection bias of long-period cometary orbits.J. Am. Stat. Assoc., 98(463):515–521

2003
[13]

Kalf, H. (1995). On the expansion of a function in terms of spherical harmonics in arbitrary dimen- sions.Bull. Belg. Math. Soc. Simon Stevin, 2(4):361–380

1995
[14]

and Jones, M

Kato, S. and Jones, M. C. (2015). A tractable and interpretable four-parameter family of unimodal distributions on the circle.Biometrika, 102(1):181–190. Klemelä, J. (2000). Estimation of densities and derivatives of densities with directional data.J. Multivar. Anal., 73(1):18–40

2015
[15]

Magnus, J. R. and Neudecker, H. (1999).Matrix Differential Calculus with Applications in Statistics and Econometrics. WileySeriesinProbabilityandStatistics.JohnWiley&Sons, Ltd., Chichester

1999
[16]

Magnus, W., Oberhettinger, F., and Soni, R. P. (1966).Formulas and Theorems for the Special Func- tions of Mathematical Physics, volume52ofDie Grundlehren der mathematischen Wissenschaften

1966
[17]

Mardia, K. V. and Jupp, P. E. (1999).Directional Statistics. Wiley Series in Probability and Statistics. Wiley, Chichester

1999
[18]

Pewsey, A. (2026). On Jeffreys’s cardioid distribution.Comput. Stat. Data Anal., 213:108248

2026
[19]

and García-Portugués, E

Pewsey, A. and García-Portugués, E. (2021). Rejoinder on: Recent advances in directional statistics. Test, 30(1):76–82

2021
[20]

Pewsey, A., Neuhäuser, M., and Ruxton, G. D. (2013).Circular Statistics in R. Oxford University

2013
[21]

van der Vaart, A

Press, Oxford. van der Vaart, A. W. (1998).Asymptotic Statistics, volume 3 ofCambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, Cambridge

1998
[22]

and Shimizu, K

Wang, M. and Shimizu, K. (2012). On applying Möbius transformation to cardioid random variables. Stat. Methodol., 9(6):604–614

2012
[23]

Wehrly, T. E. and Johnson, R. A. (1980). Bivariate models for dependence of angular observations and a related Markov process.Biometrika, 67(1):255–256. Wolfram Research, Inc. (2021). Mathematica. Computer software, Version 13.0. 53

1980

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [1]

and Capitanio, A

Azzalini, A. and Capitanio, A. (2014).The Skew-Normal and Related Families, volume 3 ofInstitute of Mathematical Statistics Monographs. Cambridge University Press, Cambridge

2014

[2] [2]

and Biswas, S

Banerjee, B. and Biswas, S. (2024). Intrinsic geometry-inspired dependent toroidal distribution: Application to regression model for astigmatism data.arXiv preprint, arXiv:2409.06229

arXiv 2024

[3] [3]

and Grübel, R

Baringhaus, L. and Grübel, R. (2024). Discrete mixture representations of spherical distributions. Stat. Pap., 65(2):557–596

2024

[4] [4]

and Ebner, B

Borodavka, J. and Ebner, B. (2026). A general maximal projection approach to uniformity testing on the hypersphere.Bernoulli, to appear. Chacón, J. E. and Duong, T. (2018).Multivariate Kernel Smoothing and its Applications, volume 160 ofMonographs on Statistics and Applied Probability. CRC Press, Boca Raton. Chacón, J. E., García-Portugués, E., and Meilán...

2026

[5] [5]

and Xu, Y

Dai, F. and Xu, Y. (2013).Approximation Theory and Harmonic Analysis on Spheres and Balls. Springer Monographs in Mathematics. Springer, New York. 51 DLMF(2025).NIST Digital Library of Mathematical Functions. F.W.J.Olver, A.B.OldeDaalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller and B. V. Saunders, eds

2013

[6] [6]

Dones, L., Brasser, R., Kaib, N., and Rickman, H. (2015). Origin and evolution of the cometary reservoirs.Space Sci. Rev., 197(1):191–269

2015

[7] [7]

Ebner, B., Henze, N., and Meintanis, S. (2024). A unified approach to goodness-of-fit testing for spherical and hyperspherical data.Statistical Papers, 65(6):3447–3475. Fernández-de-Marcos, A. and García-Portugués, E. (2023). On new omnibus tests of uniformity on the hypersphere.Test, 32(4):1508–1529. Fernández-Durán, J. J. (2004). Circular distributions ...

2024

[8] [8]

Gradshteyn, I. S. and Ryzhik, I. M. (2014).Table of Integrals, Series, and Products. Academic

2014