Questions tagged [cumulants]
The $n$th cumulant of a random variable $X$ is the $n$th derivative of the Taylor series expansion of $\log[E(e^{tX})]$ evaluated at zero.
42
questions
0
votes
0
answers
14
views
By means of what distribution can I match the first n moments for arbitrary (i e. any) values of those moments?
Suppose I have the first n moments from some data set, either raw, centered or scaled, (or cumulants instead) whichever is more convenient for matching. Is there a continuous, continuously ...
- 3,187
1
vote
0
answers
18
views
How can I implement moment matching with kernel tricks if I do not have the complete distribution but only the higher-order moments or culuments?
As is said in Appendix B.3 of ref, "It is difficult to match high-order moments, because we have to deal with high order tensors directly. On the other hand, MMD can easily match high-order ...
3
votes
1
answer
136
views
Zero variance but non-zero skewness
I was thinking of a hypothetical distribution where the mean(first cumulant) is non-zero, second cumulant(variance) is zero, and the third cumulant(skewness) is non-zero. The higher order cumulants ...
- 133
0
votes
0
answers
49
views
How do continuous partial derivatives depend on $n$ in maximum likelihood estimation?
I'm reading Tensor Methods in Statistics by McCullagh 1987, (P209 for this question) and I can't understand one step he uses.
He begins with the usual log-likelihood
\begin{equation*}
l(\theta; Y) =...
- 41
3
votes
0
answers
123
views
Pooled Kurtosis Estimator Using Pooled Cumulant Estimators
I am trying to come up with a statistically sensible pooled kurtosis estimator that is based on pooled cumulant estimators.
Specifically, I have unbiased estimators of the second and fourth cumulant ...
- 435
2
votes
0
answers
50
views
Fenchel conjugate of the cumulant function (Exponential Family)
We have a minimal exponential family $f(x) = h(x) \exp(\langle\theta, t(x)\rangle - A(\theta)).$ The canonical parameter space is $\Theta = \{ \theta \in \mathbb{R}^d \colon \int h(x) \exp(\langle\...
- 656
0
votes
0
answers
86
views
How to call $\mathrm{Cov}(X,Y,Z)$? [duplicate]
Let
$$\mathrm{Cov}(X,Y,Z) = \mathrm E[(X-\mathrm E(X))(Y-\mathrm E(Y))(Z-\mathrm E(Z))]$$
This can be regarded as a generalization of covariance to three random variables.
I am writing a text where I ...
- 4,480
1
vote
0
answers
106
views
Does zero cumulants imply independence?
Question: Suppose we have two random variable $X$, $Y$ that follow non-Gaussian distribution, and we are given that:
$$\operatorname { cum }(X, Y)=\operatorname { cum }(X, X, Y)=\operatorname { cum }(...
- 111
0
votes
0
answers
144
views
Antiderivative of log-partition function of exponential family
Let $Y$ be a random variable with distribution belonging to a minimal regular exponential family. Let $\eta $ denote the scalar-valued canonical parameter of the exponential family and let $A$ be the ...
- 291
4
votes
1
answer
217
views
Skewness and kurtosis for $\mathrm{MA}\left(\infty\right)$ model with non-gaussian noise
If an ARMA model formulation is written in infinite moving-average form:
\begin{equation}
X_t = C\left(B\right)\epsilon_t \quad
\mbox{with} \quad C\left(B\right)=C_0+C_1B+C_2B^2 + \ldots
\end{...
- 704
1
vote
1
answer
657
views
Show that this probability distribution is an exponential family
We define a probability distribution on the non-negative integers 0,1,2,...,as having point probabilities $$P_\eta(k)=G(k,\rho)\left(\frac{1}{\rho+\eta}\right)^{\rho}\left(\frac{\eta}{\rho+ \eta}\...
- 305
1
vote
0
answers
129
views
Skewness of the sum of dependent variables [closed]
I have two random variables $X$ and $Y$ which are both time series of observations. The two series are not independent. I want to find the skewness, kurtosis, hyper-skewness and hyper-kurtosis of $X+...
- 704
2
votes
0
answers
81
views
asymptotic distribution of 3rd and 4th sample cumulants?
Suppose $X$ is distributed as standard normal, I take a sample of size $n$, and compute 3rd and 4th sample cumulants $\kappa_3$ and $\kappa_4$. I'm interested in the asymptotic distribution of $\kappa$...
- 6,237
1
vote
1
answer
1k
views
n'th cumulant (of a CGF) for exponential family / exponential dispersion model
The n'th cumulant is defined to be the n'th derivative of the CGF (cumulant generating function).
$$\kappa_n = \frac{d^n K(t)}{dt^n} |_{t=0}
$$
But I'm reading in a book (p.215, chapter5, eq. 5.8) ...
- 3,560
2
votes
1
answer
514
views
Cumulant of sum of correlated random variables?
Let $X,Y$ be two random variables. We denote by $[X^k]$ and $[Y^k]$ the $k$'th order cumulants of $X$ and $Y$, respectively. I'm interested in computing the $k$'th order cumulant of $Z = X+Y$.
If $X,Y$...
- 4,480
6
votes
2
answers
1k
views
Formulas for higher order cumulants
I want to calculate higher-order joint cumulants for 2 variables. I calculated the higher order single-variable and bivariate moments numerically. Now I need to combine them into cumulants (upto the ...
- 1,486
1
vote
0
answers
68
views
What are the explicit formulas for the cumulants in terms of z-scores?
I'm trying to calculate the first few cumulants of a random variable using $Z$-scores.
The situation
Suppose we have a random variable $X$ with mean $\mu$ and standard deviation $\sigma$, and define ...
- 237
3
votes
0
answers
37
views
Generating a 1-d sample with desired statistics
I'm interested in obtaining a sample of numbers $x_1,\ldots,x_n$ such that their cumulants approximately match user-provided set of cumulants.
For instance, I can get a sample with first two ...
- 6,237
2
votes
2
answers
1k
views
Name of third cumulant?
The first cumulant is called the mean. The second is the variance.
Does the third cumulant have a name? The fourth?
- 4,480
4
votes
1
answer
3k
views
Moments of an AR(1) Process
Definition of an AR(1) process
In an Autoregressive Process, a time series can be generated based on a stochastic difference equation.
\begin{align}
X_t = c + \phi \, X_{t-1} + \epsilon
\end{align}
...
- 592
0
votes
1
answer
97
views
What is the cumulants of a whole data in terms of the cumulants of its parts?
I have around 8 billion data points, and I need to calculate the distribution and the cumulants of this distribution.
However, due to technical restrictions, and time constraints, I can only ...
- 237
2
votes
1
answer
408
views
Expectation Value of a Product of Many IID variables
First of all, I apologize for not being rigorous, but I am not a statistitian by background.
Imagine you have $N$ i.i.d. positive random variables $X_1...X_N$ and you are trying to compute a ...
- 133
3
votes
1
answer
299
views
Cumulants of Poisson random variable conditioned on a Bernoulli random variable
Consider a Bernoulli distributed random variable $Y$, which is 1 with probability $p$ and 0 with probability $1-p$. Further there is a random variable $X$ where the conditional probability ...
2
votes
1
answer
3k
views
Higher order moments of a multivariate Gaussian rv
Let $X~N_d(\mu,\Sigma)$ be a multivariate Gaussian random vector. Is there a convenient formula for each of
$$
\mu_p\triangleq \mathbb{E}\left[\sum_{i=1}^d |X_i|^p\right],
$$
in terms of $\mu$ and ...
- 554
1
vote
0
answers
15
views
What is it called cumulative of every X previous days?
Let's say I have sales number of last 60 days.
I can just draw a graph to see how sales has changed overtime. Numbers vary in different days, like on weekends sales numbers decrease.
However, I ...
- 111
2
votes
0
answers
178
views
Are the cumulants of sufficient statistics finite for the exponential family?
Under what conditions are the cumulants of the sufficient statistic finite for an exponential family?
If we have
$$
p(x \mid \theta) = \exp(\theta \cdot T(x) - A(\theta))
$$
then the derivatives of $...
3
votes
1
answer
3k
views
Relations between moments and cumulants
From the definition of KGF (cumulant generating function) we can write:
\begin{align}
K_x(t) &= \log_eM_x(t) \\
&= \log_e\left[1+\sum_{r=1}^{\infty}\frac{t^r}{r!}\mu_r^{'}\right] \\
&= ...
- 205
3
votes
1
answer
10k
views
Handling cumulative sum of variable in time series modelling
I would like to build a prediction model, based on the time series data, but all the features are a cumulative sum over the period. For example, the value of feature x is
50 at t = 1 (original value ...
- 109
3
votes
1
answer
573
views
How can I get the cumulant expression from the recursive relation between cumulant and moment?
I am reading some paper about high-order statistics https://link.springer.com/article/10.1007%2Fs11004-009-9258-9?LI=true. The paper gives two recursive expressions relating the multivariate cumulants ...
- 543
0
votes
1
answer
1k
views
Moment generating function of the natural sufficient statistics of Gamma distribution
I have $X_1,X_2 ... X_n$ Gamma-distributed r.v with density:
First of all I showed, that Gamma distribution belongs to exponential family and can be represented in form
I found, that for Gamma ...
- 77
6
votes
1
answer
1k
views
Does exponential family of distributions have finite expected value?
I am curious about this question, because in definitions I have never seen this property. Is it true? If yes, why?
- 77
1
vote
0
answers
81
views
Saddlepoint approximation with weibull distribution
I have some trouble with this computation, I have the moment generating function of a random variable $S$ by:
$$M_S(t)=\frac{\beta\mu t}{1+(1+\beta)\mu t-M_X(t)}$$
According to the text that I am ...
- 1,003
6
votes
1
answer
1k
views
Skewness of Tweedie distribution
Tweedie distributions are a family of distributions from the exponential dispersion family that have power-law mean-variance relationship:
\begin{align}
\mathbb E[X] &= \mu \\
\operatorname{Var}[...
- 106k
1
vote
1
answer
1k
views
Expansion of Cumulant Generating Function of Negbin
Let $X \thicksim Negbin(r,p)$ where $(0\lt p \lt 1) $
I want to derive skewness and kurtosis of $X$ by getting the Cgf of X.
First, since Followance of Negative Binomial equals to the distribution ...
- 215
3
votes
0
answers
106
views
Variance of k-statistic in large-n limit
k-statistics are unbiased estimators of cumulants. In the large-n limit, the formula for the $r$th k-statistic is just the formula for the $r$th cumulant in terms of central moments. But what, in the ...
- 2,261
5
votes
2
answers
2k
views
Proof of recurrence between cumulants and central-moments
According to Wikipedia, the $n$th cumulant $\kappa_n$ is related to the central-moments $\theta_n$ by the following recurrence:
$$\kappa_n = \theta_n - \sum_{m=1}^{n-1} \binom{n-1}{m-1} \kappa_m \...
- 4,480
2
votes
3
answers
1k
views
How to obtain the probability density function from its cumulants?
Let $\sigma_1, \sigma_2, \dots, $ be the sequence of cumulants of a probability density function $p(x)$. How can we reconstruct $p(x)$ from its cumulants?
P.S. If it helps, you can assume that $p(x)$ ...
- 4,480
1
vote
0
answers
33
views
Joint cumulants of Zn2 characters
Let $f_{c}:Z_2^n \rightarrow \{-1,1\}$ be the character defined as $f_c(x) = (-1)^{<x,c>}$, where $c,x \in Z_2^n$. It is easy to see that since $f_{c_1}\cdot\ldots\cdot f_{c_k} = f_{c_1 \oplus \...
- 547
11
votes
1
answer
5k
views
Moment Generating Functions and Fourier Transforms?
Is a moment-generating function a Fourier transform of a probability density function?
In other words, is a moment generating function just the spectral resolution of a probability density ...
- 210
1
vote
0
answers
259
views
Relative influences of terms in the law of total cumulance
I posted this to mathoverflow and no one's saying anything.
$\newcommand{\cum}{\operatorname{cum}}$
Denote the joint cumulant of several random variables by $\cum(A,B,C,\ldots)$ (more precisely, the ...
- 10.7k
5
votes
1
answer
4k
views
Difference between cumulants and moments
In particular, is the $n$th cumulant equivalent to the $n$th central moment (i.e. about the mean)?
There's little difference I can see between MGFs (moment generating) and CGFs (cumulant generating), ...
- 790
16
votes
1
answer
705
views
Distribution with $n$th cumulant given by $\frac 1 n$?
Is there any information out there about the distribution whose $n$th cumulant is given by $\frac 1 n$? The cumulant-generating function is of the form
$$
\kappa(t) = \int_0 ^ 1 \frac{e^{tx} - 1}{x} \...
- 9,012