Questions tagged [numerics]
Also known as Numerical Analysis, Numerics aims to provide methods and algorithms for numerical computations.
1
vote
0answers
41 views
Bayesian Networks - Factor Graphs - Belief Propagation - Numerical stability
I am trying to do inference for a Bayesian Network with discrete probabilities.
I converted the network to a factor graph and implemented the sum-product algorithm (belief propagation). My goal is ...
1
vote
1answer
39 views
Given unnormalized weights in logs, how do I compute normalized weights in logs?
Setting
Given a set of positive weights $\{w_i\}_{i=1}^n$, I can normalize them by computing
$$W_i = \frac{w_i}{\sum_{j=1}^nw_j}\quad \forall i=1,...,n.$$
Easy enough. But for numerical reasons, it ...
0
votes
0answers
15 views
Example of method of moments solved numerically?
I have come across many examples where the method of moments estimators can be found in closed form (Exponential, Normal, and etcetera). However, there are cases where there is no closed form moments, ...
0
votes
0answers
7 views
how to systematically choose the thresholding in least square solver, for solving ill-posed least square problem?
I have a system
$Ax =b$, where $A\in\mathbb{R}^{300\times 200}$, but $rank(A)=70$.
And I know the true solution.
I tried standard least square solver such as ...
-1
votes
0answers
14 views
Easy way to sample a Bayes posterior distribution of stable distributions?
I have a markov chain $P(x_{i+1}|x_i)=\rho(x_{i+1} ; \alpha,\beta,c, x_i)$, where $\rho$ is the stable distribution with mean $x_i$. I'm interested in fixing $x_1$ and $x_3$, and sampling an $x_2$ ...
1
vote
0answers
44 views
Natural Splines and Smoother Matrix
In the context of smoothing splines, one can show that the Reinsch form is given by:
$
\hat{y} = N (N^{T}N +\lambda \Omega)^{-1}N^{T} y
= (I+ \lambda K)^{-1}y
$
where (1) $K = (N^{T})^{-1}\Omega N^{-...
1
vote
0answers
49 views
Moment generating function of a Weibull distribution and root finding heavy and light tailed case
I consider the equation $M_x(v)=1+(1+\beta)\mu$ and I need to find the solution $v>0$ such that the equation is fulfilled.
For this example I consider the moment generating function $M_X(v)$ of a ...
0
votes
0answers
16 views
Optimize function on two stages
I have a non-linear cost function (which I do not know its analytical derivative and it is pretty hard to be computed). The function takes 8 parameters.
I need to minimize this function using some ...
0
votes
0answers
54 views
Estimating correlation matrix using numeric likelihood maximization
I'm performing maximum likelihood estimation on jointly distributed data and I'm having some issues estimating the correlation terms. I am using an approach based on the Cholesky decomposition, but I ...
3
votes
1answer
25 views
Are analytical derivatives unambiguously superior to numerical derivatives in GMM?
I am estimating a non-linear GMM model. In both Stata and R, you need to specify the moment equations and the instruments, but there is no need need to provide analytical derivatives for the estimator ...
1
vote
0answers
30 views
Underflow when estimating marginal likelihood via bridge sampling
I try to use an iterative procedure to estimate the marginal likelihood in a Bayesian setting for model selection. In case you are interested in the specifics of bridge sampling in my application, see ...
3
votes
0answers
41 views
Optimization with/without an analytical gradient
A colleague is optimizing a function (e.g. trying to find the minimum of a function $f(x_1, x_2, \ldots)$). We know the analytical form and it is differentiable. I suggested calculating the ...
0
votes
0answers
41 views
Numerical Solution for Logistic Regression
Reading the Numba tutorial, I found a neat numerical solution for logistic regression
...
2
votes
1answer
135 views
Are mvrnorm() in MASS R package and rmvn() in mgcv R package equivalent?
I am carrying out posterior simulation with GAMs/SCAMs and was wondering if/how the rmvn() function differs in any way from the ...
0
votes
0answers
17 views
Discretize numeric attribute values
I have a numeric attribute (e.g. revenue attribute), I must discretize its values before I use WEKA tool, but based on what I must divide them into categories? Since this may affect the classification ...
1
vote
1answer
26 views
numberical implementation of linear regression with “loose variable”
I understand how to solve a linear system $X \beta = y$
the solution is
$\beta = (X^{T}X)^{-1} X^{T} y$
The problem is I could have an entry $\beta_i$ where it has no exposure in $X$. i.e. $X$ has ...
3
votes
0answers
26 views
An urn problem: few red balls, many draws (with replacement)
So, this is a freshman probability problem and I am embarrassed to p[ost it, but I have been up for 35 hours and my brain is broken.
I have an urn with 60,000 white balls and 6 red balls. From this ...
2
votes
0answers
43 views
Dealing with Numerical Issues from Computing Weighted Sample Covariance Matrix
I have vector-valued samples $\mathbf{x}_1, \mathbf{x}_2, \ldots, \mathbf{x}_N$ with normalized weights $w_1, \ldots, w_N$. I'm trying to compute a weighted sample covariance matrix:
$$
\sum_i w_i (\...
1
vote
0answers
40 views
How many possibilities to divide 12 elements into 4 buckets of 3 elements [closed]
Say we have
possibilities: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 and 12
4 buckets that can contain 3 elements
How may possibilities are there to divide the 12 elements over the 4 buckets (where each ...
0
votes
0answers
362 views
Formula for Inverse T-distribution
I am trying to formulate an expression to calculate the critical value of a T-distribution for a given degrees of freedom. I have done so already for the Normal Distribution by considering the TI-84 ...
5
votes
2answers
131 views
Parameter estimation without an explicit likelihood function
I have a parametric model, some data $y$, and I would like to find a maximum likelihood estimate for the model parameters $\theta$. My usual approach would be to write down the likelihood function $\...
4
votes
0answers
156 views
Choosing the basis functions in a linear regression
I have two random variables $X$ and $Y$ and I'm trying to model $\mathbb{E}[Y|X]$.
To this end, I'd like to pick a collection of functions $f_1, f_2 \dots f_n : \mathbb{R} \to \mathbb{R}$ and then ...
1
vote
0answers
455 views
Deep Learning: Condition Number and Poor Conditioning
I am reading the following section of the book Deep Learning.
Can you provide an intuitive explanation of the above section? I don't quite understand the statement "When this number is large, matrix ...
2
votes
0answers
50 views
Training an Artificial Neural Network with limited-memory Quasi-Newton
I would like to train a simple Artificial Neural Network implementing an algorithm of the class of limited-memory Quasi-Newton.
I read the paper Modified quasi-Newton methods for training neural ...
4
votes
2answers
125 views
Why is two-sided gradient checking more accurate? [closed]
In week 5 of Andrew Ng's Machine Learning course, he gives the formulae for gradient checking:
One-sided difference:
$\dfrac{\partial}{\partial\Theta}J(\Theta) \approx \dfrac{J(\Theta + \epsilon) - ...
0
votes
0answers
60 views
Integral of probit likelihood
I'm currently trying to implement the Heckman method for estimating the dynamic probit panel model (Original paper can be found here). I'm trying to implement it according a paper of Stewart (2006), ...
1
vote
0answers
242 views
GMM Estimation and convergence problem
I try to minimize an unweighted moment function $G(\theta)$ given by $G(\theta) = \bar{g}(\theta)'\bar{g}(\theta) $.
$g(\theta,x_i)$ contains the specified moment conditions, where we state $E(g(\...
2
votes
1answer
227 views
Moments of the truncated normal distribution (univariate) away from the mean
I need to compute the mean and variance of the truncated normal distribution. For simplicity, let us focus on a standard normal, since the general case can be reduced to this. The PDF is given by:
$$...
1
vote
0answers
60 views
Is there a solution for Canonical Correlation Analysis on large sparse matrices?
I'm trying to run CCA over two views which are sparse matrices. The two views are very high dimensional (e.g. 300k, 400k) with 1m samples.
CCA needs the input views to be zero mean but I won't be ...
24
votes
4answers
8k views
Why does Andrew Ng prefer to use SVD and not EIG of covariance matrix to do PCA?
I am studying PCA from Andrew Ng's Coursera course and other materials. In the Stanford NLP course cs224n's first assignment, and in the lecture video from Andrew Ng, they do singular value ...
0
votes
1answer
48 views
Distribution of infinite sum $\sum_{t=0}^{\infty} \epsilon_t r^t $
In my current statistics course we're being taught about time series, and in this context we came across sums like this:
$$\sum_{t=1}^{\infty} \epsilon_t r^t \quad \epsilon_t\sim \text{WN}(0,\sigma^2) ...
0
votes
1answer
93 views
Method to compute empirical derivative about some point
I have black-box access to some function and I want to compute the derivative about the point X. Is there a method that does this?
4
votes
1answer
2k views
Softmax overflow [closed]
Waiting the next course of Andrew Ng on Coursera, I'm trying to program on Python a classifier with the softmax function on the last layer to have the different probabilities.
However, when I try to ...
4
votes
1answer
510 views
What is the best way to apply the log-sum-exp trick in this situation?
I am aware of the "log-sum-exp" trick for calculating the logarithm of sums that handles overflow and underflow issues. However, I would like to know more about how it works. In particular, I am ...
1
vote
1answer
41 views
Numerical method to find x [closed]
I am carrying out a simulation study using the cumulative distribution function $$ F(x)=\frac{4x\arctan\frac{x}{a}}{a^2(π-2)}$$
Now I need to get x subject in order to carry out the simulation study. ...
1
vote
0answers
46 views
Proving an equivalence relation by using numerical methods such as Gaussian quadrature
The background is residual useful life prediction. The following is my problem description.
Degradation signal path: $r(t)=\phi+\theta t$ , where $\phi$ is assumed to be the same for all units，and $\...
4
votes
0answers
144 views
What is the fastest way to compute PC1 scores, without performing the whole PCA?
I want to compute only the first principal component's scores $t_1$ of a large number $n$ of data points x with a high dimensionality $p$. Assume the data has been centered about zero. Data points ...
2
votes
0answers
340 views
Comparison between analytic gradient and numerical gradient for multivariate normal distribution wrt mean and covariance
The analytic gradients of log multivariate normal distribution wrt mean and covariance matrix can be found at StackExchange post and The gradient of the log-likelihood of normal distributions.
I ...
6
votes
1answer
203 views
Relationship between binomial regression link function and goodness-of-fit tests [now with link to R code]
Some background: A number of papers in the literature (various ones by Hosmer and Lemeshow; Copas; le Cessie and van Houwelingen; Cressie and Read; Osius and Rojek; J. R. Dale) discuss a family of ...
2
votes
0answers
689 views
Logistic regression and singular Hessian
I've been following along with Andrew Ng's excellent course on Machine Learning (CS 229), and have been working on Problem Set 1.
I'm trying to fit a logistic regression model using Newton's Method. ...
0
votes
0answers
53 views
Accurate tail probabilities (p-values) in MASS package
An R package was reporting a p-value for my dataset to be zero. Digging into it, I found that the p-values were computed in MASS:::summary.loglm :
...
7
votes
2answers
981 views
Is it possible to have pearson correlation coefficient values < -1 or values > 1?
I am trying to calculate the Pearson correlation coefficient according this formula over a large dataset:
Mostly, my values are between -1 and 1, but sometimes I get weird numbers like:
...
3
votes
1answer
918 views
Why do prcomp() and eigen(cov()) in R return different signs of PCA eigenvectors?
I understand the sign of the eigen vectors / PCA rotations can be positive or negative (see here or here).
But I am curious why the following two approaches yield different results, from numerical ...
1
vote
0answers
23 views
What analysis can be performed for categorical and numeric factors and their interactions on numeric response variable?
What analysis can be performed to see the effects of categorical and numeric factors and their interactions on a continuous or numeric response variable? It seems that I need to use ANCOVA, but not ...
-2
votes
2answers
203 views
plotting a factorial function in R
I have problem trying to plot the following function in R:
The R code I'm using is:
...
1
vote
1answer
192 views
Interpolate a CDF to get an interpolated hazard rate, or interpolate the hazard rate directly?
My problem is that I need to do an interpolation. Eventually, I will work on the hazard rate, but I do not know if it is better to interpolate the CDF or the hazard rate. Let me explain better.
I've ...
2
votes
1answer
273 views
Monotonicity of special case of Kullback-Leibler divergence
I have two discrete distributions $\tau$ and $\rho$ with the same support $\Omega$. I'm considering a weighted mixture of these distributions described by the following function:
$$
f(w) = (1-w) \cdot ...
5
votes
2answers
4k views
Kullback-Leibler Divergence
I tried to implement a numerical estimate of the Kullback-Leibler Divergence for two samples. To debug the implementation draw the samples from two normal distributions $\mathcal N (0,1)$ and $\...
4
votes
2answers
151 views
Does rounding introduce variance into estimates?
It is often recommended to round parameter estimates to avoid suggesting more precision than the data really have, e.g. here. I understand rounding does not introduce bias, as long as an unbiased ...
0
votes
0answers
35 views
Good algorithm to compute the sample mean and variance [duplicate]
I am looking for good (fast, memory efficient, ideally one-pass) algorithms to compute the sample mean and variance.
I have found a technical report by Chan, Golub, and LeVeque from 1983 with ...
