Questions tagged [fourier-transform]
The Fourier transform decomposes a signal (a function of time) into frequencies, giving the energy at each frequency.
167
questions
0
votes
0answers
16 views
Difference of two independent lognormal random variables by Fourier transform methods?
Is it possible to calculate the difference $X_1-X_2$ of two independent lognormal variables $X_1$ and $X_2$ where $\log(X_1)\sim N(\mu_1,\sigma_1)$ and $\log(X_2)\sim N(\mu_2,\sigma_2)$? Could I ...
3
votes
1answer
47 views
Characteristic function and Fourier transform for a discrete random variable!
Let $\phi_{x}(t)= E [ e^{itx}]$ be the characteristic function
If X is a continuous random variable, then:
$\phi_{x}(t)= E [ e^{itx}] = \int e^{itx} f(x)dx$ (being $f(x)$ the probability density ...
0
votes
0answers
15 views
Method to calculate optimal nbasis (K) for fourier basis
Is there any method to calculate the optimal K for fourier basis transformation?
For example if I were to use RMSE, it seems that as K increases RMSE keeps going lower.
I'm thinking of the elbow-...
1
vote
1answer
25 views
Estimating Fourier spectrum from multiple time series of a system
I have a set of N time series, each of length T, that describe separate realisations of a single physical system. For each series, I can compute an FFT to find the Fourier spectrum up to a period 2/T, ...
0
votes
0answers
21 views
Long term time-series prediction for non-stationary but regular water flow data
I am looking at a very interesting time series dataset for the volume of water flow through some groundwater sensors. Our goal is longer term prediction of water levels, such as 1 - 5 years in the ...
1
vote
0answers
14 views
Spectral analysis in R, how to interpret periodogram
I'm relatively new to spectral analysis and have been working through some online tutorials. I have some time series data that I would like to examine for periodicity / repeating patterns.
When I ...
0
votes
0answers
6 views
Can you perform a Fourier transform of a Frequency / Time plot? (Number of phone taps (t))
I'm working on a thesis problem. I need to analyze tapping behaviour (number of phone taps). The raw data is the time stamp of every app, and the app in use.
It was recommended to me that I create a ...
0
votes
0answers
12 views
Representing a time-series smoothed curve as a sinsoidal?
So attaches is an example of the kind of time series data I am working with. So far I have used Gaussian filters with sigma=3 and 6 to smooth the data, which has worked very well (especially sigma=3).
...
0
votes
0answers
9 views
Fourier transforms for noise reduction
Given a signal, which is regularly sampled over time and is noisy. The standard method is with a Fourier transform to reduce the noise and minimise the change to the signal.
...
1
vote
1answer
44 views
Why SARIMA has better accuracy on weekly dataset than on daily one?
I am studying time series right now. So, I have this dataset. My aim is temperature prediction.
I've found out that ARIMA can't work with long period seasonality. So, I've resampled daily dataset ...
0
votes
1answer
15 views
Fourier analysis to retrieve components of individual spectra
I have a basic, simple question, I am a physics student, and searching internet gives me a lot of signal processing theory but couldn't find this basic answer, which I plan to implement in my speech ...
1
vote
1answer
23 views
Log-likelihood function for a filtered Fourier spectrum
I have time series data from which I am trying to infer parameters using MCMC. I normally infer parameters about the data in the time domain, using a Normal log-likelihood. However, I now have to ...
0
votes
0answers
14 views
Exactly how are cyclical components computed?
So suppose we have some sort of a time series model.
y_t = trend_t + cyclical_t + x_t + epsilon_t
So, I'm interested in obtaining the seasonal component. Here "x_t" refers to other potential ...
0
votes
0answers
33 views
FFT for a binary time-series
I have a multivariate time-series of a binary values where 0 means that some state was not observed in the given second and 1 meaning that it was observed.
I know that for a time-series consisting of ...
2
votes
0answers
155 views
Measuring weather impact on sales as a crossed random effect
I'm trying to model sales of a clothing brand having longitudinal data (unbalanced panel: 20 stores with 50 - 157 weeks of datapoints).
There are many regressors in my analysis, but I believe that ...
0
votes
0answers
12 views
Discrete Fourier transform (DFT) of 2 signals using a single DFT
Given two different(x,y) and independent signals, can one find dft of both of them using a single dft chip and using that that chip only once. I tried finding dft of x+jy but was not able to find ...
3
votes
1answer
27 views
Periodicity of random kitchen sink feature mappings
In various papers, e.g. Random Features for Large-Scale Kernel Machines, Rahimi and Recht introduce the now popular methodology wherein a "low rank" approximation to a stationary, PSD, kernel $K(x,y) =...
0
votes
0answers
23 views
Relationship between multivariate normal random variable and its Fourier transform
I was looking at the distribution of N dimensional real momenta $\pi$ where
$\pi \sim \mathcal{N}(0,M)$. It turns out that $M$ is diagonal (and real) in Fourier space and so I wanted to transform to ...
0
votes
0answers
5 views
Why assumption of negligible cross-correlation between missing points is important for PSD estimate
In this paper they propose a method for estimating the Fourier transform (FT), and therefore the power spectral density of a time series with one or more missing points. They do this by inserting ...
1
vote
1answer
21 views
Fourier series question I need help with
For the case of $a_n$ I am confused how my notes say we have 1 for the case n = 0. When calculating this integral for sin(nx) when n=0, we should also get zero as at x=0, sin(x) is 0. I must be ...
2
votes
0answers
25 views
How to efficiently find spikes in 2D data projected to 1 axis
I have the following points:
My goal is to find the be able to differentiate the yellow points from the purple ones.
We can assume that the yellow points are aligned on the vertical axis.
We cannot ...
0
votes
0answers
16 views
applying fast Fourier transform without restriction on N
I want to apply fast Fourier transform for a sequence of complex numbers with length N. N can be anything (not necessarily a power of 2). It seems that Cooley–Tukey algorithm only works for the ...
0
votes
0answers
14 views
Multitaper F statistics
I'm having some problems interpreting the F-statistics output from multitaper analysis. To illustrate, the following code-snip in R performs multitaper analysis on the same sine-frequency but with ...
0
votes
0answers
51 views
extracting seasonality - using fourier transform vs. learning the coefficients of fourier terms
I'm following hyndman's advice for using fourier terms when fitting a linear regression model to the taylor time-series (with the very long seasonality of 336). My ...
1
vote
1answer
38 views
How to discard the first spike after auto-correlation and handle sloping auto-correlation output [closed]
Disclaimer: I am not very mathematically inclined and am mostly looking to be pointed in the right direction.
I have various signals that I am putting through an auto-correlation function that uses ...
0
votes
0answers
36 views
Capturing cycles in time series with Fourier
In time series regression with Fourier, Fourier terms are limited to maximum half of the frequency of time series to capture seasonality.
my question is: Can we extend Fourier terms beyond half of ...
0
votes
0answers
11 views
Marginalisation when conditioning on the Fourier transform of a random variable
I am looking to sample from a distribution $p(y)$ defined by the following expectation:
$$p(y) = \mathbb{E}_{p(u)} \left[ p(y|u) \right]$$
Both $p(u)$ and $p(y|u)$ are multivariate Gaussians:
$$p(u)...
0
votes
0answers
57 views
How to search for irregular signals: Fourier, DWT or k-means?
See my notebook here
I want to search for irregular time signals in a data set of ~3 500 000 time signals. I can't give a clear definition of irregular signal, but it must fulfil the criteria of: not ...
0
votes
1answer
34 views
What is the difference between machine learning approaches and Fourier series to fit a curve to data graph?
As I know machine learning(at least in some problems) tries to fit a curve to data graph. And I think Fourier transform tried to do it. But machine learning use a hypothesis curve with the formula ...
1
vote
0answers
173 views
Fast Non-uniform DFT in R
So, if I want to compute discrete Fourier transform (DFT) in R, I can create my own functions like so:
...
1
vote
1answer
72 views
Why is spectral density only defined for stationary processes?
I read Brockwell and Davis(2016), Shumway and Stoffer(2016), and Stoica and Moses(2004). However, none of them laid out clearly the reasoning behind the presumption of stationarity when conducting ...
2
votes
0answers
25 views
Probability of sparse spectrum
Consider a vector $v$ such that $v \sim \mathrm{Unif}(\mathbb{S}^{d-1})$, the uniform distribution on the unit sphere in $d$ dimensions.
Question: is there an upper bound on the probability that $v$ ...
0
votes
1answer
414 views
K in Fourier series - How to find value of K to use it in ARIMA?
I am using the forecast library for doing some time series forecasting. I need to forecast number of sold items. I am planning to add holidays as xreg in auto.arima. The holiday will be a 0/1 list, ...
3
votes
1answer
147 views
Continuous time Fourier representation
I have learned that the Fourier transform of a continuous-time unit-periodic stochastic process is:
$$x(t) = \sum\limits_{k=-\infty}^{\infty} a_k e^{i2\pi kt} \quad \quad \text{ where } \quad \quad ...
2
votes
2answers
327 views
Inverse Fast Fourier Transform in R
plot(fft(fft(1:100),inverse = "true")/100)
plot(1:100)
In the above example, I was expecting the plots to be identical.
The first, however, returns:
while the ...
2
votes
0answers
115 views
Frequency analysis of categorical/binary data
I want to do frequency analysis with a data set of consecutive binary values, such as "Rainy - Sunny - Rainy - Rainy - Rainy - Sunny - ...". Using this data, I want to extract the frequency (= the ...
0
votes
1answer
272 views
Maximum number of Fourier terms in forecast package
I am using the forecast package in R to get some Fourier components - namely, function fourier(ts, K, ..). For a time series <...
2
votes
0answers
59 views
Adaptive knot selection for B-spline fitting
When fitting a B-spline for regression purposes I've seen a lot of cases where knots are fixed uniformly ,but in some situations this could lead to poor estimations because the behaviour of the curve ...
2
votes
0answers
33 views
Non-linear regression. Obtain B.spline coefficients using Fourier Transform?
I came up with a idea to estimate the coefficients of a B-spline fit by using the Fourier Transform but I don't know if it makes any sense to estimate them in this way.
Given that
$$s(x)=\sum_kc(k)\...
2
votes
0answers
131 views
How to fit a model to both daily and weekly periodic data using linear regression and fourier transform
I have a variable, y, in 30 minute intervals, and is highly dependent both on the time of day and the day of the week - mainly Sundays. It is also dependent on bank holidays, which could by any day of ...
2
votes
1answer
375 views
R forecast multiple seasonality optimal model search using fourier and msts objects
Hi I have hourly data (one obs one hour) with multiple seasonality. I would like to fit an ARIMA model using forecast R package taking into account the multiple seasonality, maybe taking also in ...
1
vote
0answers
133 views
Fourier Output Meaning
I just ran a fourier series on weekly sales data for 3 years worth of data.
I optimally chose the number of k-terms based on the AIC.
First 6 lines of my data:
...
2
votes
0answers
184 views
RKHS norm and Fourier transform link
In the notes here, it is stated that norms of some reproducing kernel Hilbert spaces can be written in terms of Fourier transforms, and this is often used to argue that a higher RKHS norm implies a ...
1
vote
0answers
19 views
Is there a concept of weakly periodic in stats? If so, what is it called?
Ok, say we have a signal that reoccurs within every T+x time where T is a constant and x is a random number small in size compared to T. We know that a function is actually periodic if F(x) = F(x+T), ...
0
votes
0answers
116 views
Need help understanding output of a periodogram
In my effort to understand the output of a periodogram I created a series (s) where 1,1,1,1,1,1,1,1,1,10 is repeated 100 times and then created a periodogram of this series using the following R code:
...
0
votes
1answer
384 views
Discrete Fourier transform of an exponential decay
I have a vector with an exponential decay signal, using Numpy:
t=np.arange(128)
a=0.1
decay=np.exp(-a*t)
I would like to compute the discrete Fourier transform (...
3
votes
1answer
5k views
Forecasting with ARIMA ( Training and Test Data split)
I have an hourly time series of the average parking occupancy with data available from September 2017 up until June 2018. I would like to use the ARIMA model with external regressors to produce a ...
4
votes
0answers
821 views
Fourier transform of a Gaussian process
I would like to discuss and ask a question regarding the Fourier transform of a Gaussian process, if it makes sense.
For that purpose, let me describe the following situation.
Let $z(s)$ be a ...
3
votes
1answer
315 views
Sampling from characteristic/moment generating function
Suppose I am given a probability distribution only via its characteristic or moment generating function and I want to sample from that distribution to generate paths in a Monte Carlo simulation. Is ...
3
votes
1answer
132 views
Fourier Transform based imputation
Fourier Transform based imputation
Can any body please assist me in understanding the Fourier Transform based imputation algorithm shown in Figure.
I am struggling to understand ts and te.
Link of ...
