RockScience
  • Member for 11 years, 3 months
  • Last seen more than 1 year ago
Criteria to set STL s.window width
Accepted answer
14 votes

The question is not about whether it is a monthly or a weekly data, but about how quickly the seasonality evolves. If you think the seasonal pattern is constant through time, you should set this ...

View answer
Why square the difference instead of taking the absolute value in standard deviation?
11 votes

Estimating the standard deviation of a distribution requires to choose a distance. Any of the following distance can be used: $$d_n((X)_{i=1,\ldots,I},\mu)=\left(\sum | X-\mu|^n\right)^{1/n}$$ We ...

View answer
Mean absolute deviation vs. standard deviation
11 votes

Both measure the dispersion of your data by computing the distance of the data to its mean. the mean absolute deviation is using norm L1 (it is also called Manhattan distance or rectilinear distance) ...

View answer
Resources for learning about spurious time series regression
Accepted answer
11 votes

These concepts have been created to deal with regressions (for instance correlation) between non stationary series. Clive Granger is the key author you should read. Cointegration has been introduced ...

View answer
How to use DLM with Kalman filtering for forecasting
8 votes

I suggest you read the dlm vignette http://cran.r-project.org/web/packages/dlm/vignettes/dlm.pdf especially the chapter 3.3

View answer
What is the best software for time series analysis and forecasting?
6 votes

You can do roughly the same with the 2. Matlab is maybe cleaner and easier to use as you have only one clean library of function for each task, R is maybe more flexible as you have a LOT of ...

View answer
Stacked bar plot
6 votes

For the coloring, either you specify a list of colors or you generate them. In the latter, I suggest you execute this code n = 32; main.name = paste("color palettes; n=",n) ch.col = c("rainbow(n, ...

View answer
Maximizing probability of winning on loaded coin
6 votes

1/ do you already know the bias n? 2/ if yes (actually you need only to know which side of the coin is heavier) then you cannot do better than always bet on this side. In the long term, you'll have ...

View answer
Choosing seasonal decomposition method
5 votes

That's an answer for question 2. STL: http://www.wessa.net/download/stl.pdf X-12-ARIMA (and much more): http://www.census.gov/srd/www/sapaper/sapaper.html

View answer
What are common statistical sins?
5 votes

I would say, doing tests and regressions on a small set of data. Edit: Without looking at the confidence intervals, or when the confidence intervals/error bars are not easy to calculate.

View answer
What is the Unscented Kalman Filter?
Accepted answer
4 votes

The Unscented Kalman Filter is a type of non linear Kalman filter. (ie when the transition and observation functions are non linear) If these functions are differentiable, one can simply use the ...

View answer
Determining the nature of noise
4 votes

Do you know where does the noise comes from? Before doing any statistical test, you think about the origin of the noise you want to remove. Additive noise is independent from the level of the signal, ...

View answer
Rewriting AR model in State-Space form
Accepted answer
4 votes

I suggest you buy the excellent book by G. Petris, S. Petrone and P. Campagnoli Dynamic Linear Models with R. You will learn that any ARMA model $Y_t = \sum_{j=1}^{r}\phi_jY_{t-j} + \sum_{j=1}^{r-1}...

View answer
Mean of a sliding window in R
4 votes

library(zoo) x=c(4, 5, 7, 3, 9, 8) rollmean(x,3) or library(TTR) x=c(4, 5, 7, 3, 9, 8) SMA(x,3)

View answer
Computing correlation (and the significance of said correlation) between a pair of time series
4 votes

How do you define correlation for non stationary time series? Do you plan to take the correlation of the diff or these time series? If not, I suggest you look for cointegration rather than correlation ...

View answer
Getting started with time series in R
Accepted answer
3 votes

It seems like you need the package xts. Create your time serie using install.packages('xts') library(xts) X = xts(coredata(DF[,2]), order.by=DF[,1]) Then you will be able to manipulate your data ...

View answer
Detrending method for nonstationary data
2 votes

Detrending can be done by applying a low pass filter that calculates the trend. The remaining part is your detrended data. Two examples of low pass filters here: apply a Hodrick-Prescott filter. I ...

View answer
Interpretation of an integrable time series of an order zero
Accepted answer
2 votes

Commonly, a time series is said to be $I(0)$ if the time series itself is stationary (no need to differentiate it to obtain stationarity). The Wikipedia page you mention says that not all $I(0)$ ...

View answer
Good GUI for R suitable for a beginner wanting to learn programming in R?
2 votes

I recommend Tinn-R (Which is the acronym for Tinn is not Notepad)

View answer
What is the difference between a one-sided filter and a two-sided filter when looking at time series analysis?
1 votes

In the context of time series (but why would you use one side filters otherwise?): One side filters: because they only use past data, these filters can be used for backtesting and for online analysis ...

View answer
symmetry of linear regression
1 votes

Package MethComp has a function Deming which is performing this regression http://cran.r-project.org/web/packages/MethComp/MethComp.pdf The Deming regression is a special example of the total least ...

View answer
What to do when the intersection of some time series doesn't contain enough data point?
Accepted answer
1 votes

If you use R you can maybe 1/ merge the 2 time series 2/ carry forward the values except if the delay is too long (kind of enhanced na.locf)

View answer
Software for easy-yet-robust data exploration
1 votes

In my opinion, if you don't code yourself the test, you are prone to errors and misunderstandings of the results. I think that you should recommend them to hire a statistician that has computer ...

View answer
Is there a site to post my survey to so I can get a sample representative of the population?
1 votes

a sample representative of population cannot be obtained through internet as you will only get people interested in answering your survey online, which will give you a biased sample.

View answer
Chi-square analog for context-dependent distributions
1 votes

Sounds like you'll need a HMM to do that. Have you read Lawrence R. Rabiner (February 1989). "A tutorial on Hidden Markov Models and selected applications in speech recognition" There are a few ...

View answer
Is the Lagrange function objective plus lambda times constraints or objective minus lambda times constraints?
1 votes

It is exactly the same! You want the constraint to be respected, and you don't care about the sign of g(x,y)

View answer
Modeling of real-time streaming data?
1 votes

Bayesian networks are perfect for online estimation, and offer a great diversity of models.

View answer
Book for broad and conceptual overview of statistical methods
0 votes

I would recommend "Time Series Analysis and its applications with R examples" by Shumway and Stoffer The third edition: http://www.stat.pitt.edu/stoffer/tsa3/ Click and buy http://www.amazon.com/...

View answer
Outlier detection for generic time series
0 votes

Winsorization replaces extreme data values with less extreme values. http://www.r-bloggers.com/winsorization/

View answer
Testing paired frequencies for independence
0 votes

I would maybe start with a rank correlation analysis. The issue is that you may have very low correlations as the effects you are trying to capture are small. Both Kendall and Spearman correlation ...

View answer